diff --git a/nb/HuggingFace Course-Llama3.1_(8B)-GRPO.ipynb b/nb/HuggingFace Course-Llama3.1_(8B)-GRPO.ipynb index 7705a50..4499413 100644 --- a/nb/HuggingFace Course-Llama3.1_(8B)-GRPO.ipynb +++ b/nb/HuggingFace Course-Llama3.1_(8B)-GRPO.ipynb @@ -693,14 +693,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", diff --git a/nb/HuggingFace Course-Mistral_v0.3_(7B)-GRPO.ipynb b/nb/HuggingFace Course-Mistral_v0.3_(7B)-GRPO.ipynb index 46a959f..8c65665 100644 --- a/nb/HuggingFace Course-Mistral_v0.3_(7B)-GRPO.ipynb +++ b/nb/HuggingFace Course-Mistral_v0.3_(7B)-GRPO.ipynb @@ -545,14 +545,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", diff --git a/nb/HuggingFace Course-Phi_4_(14B)-GRPO.ipynb b/nb/HuggingFace Course-Phi_4_(14B)-GRPO.ipynb index 54a68e4..2b6b1c1 100644 --- a/nb/HuggingFace Course-Phi_4_(14B)-GRPO.ipynb +++ b/nb/HuggingFace Course-Phi_4_(14B)-GRPO.ipynb @@ -317,14 +317,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", diff --git a/nb/HuggingFace Course-Qwen2.5_(3B)-GRPO.ipynb b/nb/HuggingFace Course-Qwen2.5_(3B)-GRPO.ipynb index 19af88b..5831c8b 100644 --- a/nb/HuggingFace Course-Qwen2.5_(3B)-GRPO.ipynb +++ b/nb/HuggingFace Course-Qwen2.5_(3B)-GRPO.ipynb @@ -834,14 +834,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", diff --git a/nb/Kaggle-Llama3.1_(8B)-GRPO.ipynb b/nb/Kaggle-Llama3.1_(8B)-GRPO.ipynb index d014c5f..ed305f7 100644 --- a/nb/Kaggle-Llama3.1_(8B)-GRPO.ipynb +++ b/nb/Kaggle-Llama3.1_(8B)-GRPO.ipynb @@ -8,7 +8,7 @@ "
\n", "\n", "\n", - " Join Discord if you need help + \u2b50 Star us on Github \u2b50\n", + " Join Discord if you need help + ⭐ Star us on Github ⭐\n", "
\n", "\n", "To install Unsloth on your own computer, follow the installation instructions on our Github page [here](https://docs.unsloth.ai/get-started/installing-+-updating).\n", @@ -48,7 +48,19 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\n!pip install --upgrade -qqq uv\ntry: import numpy; get_numpy = f\"numpy=={numpy.__version__}\"\nexcept: get_numpy = \"numpy\"\ntry: import subprocess; is_t4 = \"Tesla T4\" in str(subprocess.check_output([\"nvidia-smi\"]))\nexcept: is_t4 = False\nget_vllm, get_triton = (\"vllm==0.10.1\", \"triton==3.2.0\") if is_t4 else (\"vllm\", \"triton\")\n!uv pip install -qqq --upgrade unsloth {get_vllm} {get_numpy} torchvision bitsandbytes xformers\n!uv pip install -qqq {get_triton}\n!uv pip install \"huggingface_hub>=0.34.0\" \"datasets>=3.4.1,<4.0.\n!uv pip install transformers==4.55.4" + "source": [ + "%%capture\n", + "!pip install --upgrade -qqq uv\n", + "try: import numpy; get_numpy = f\"numpy=={numpy.__version__}\"\n", + "except: get_numpy = \"numpy\"\n", + "try: import subprocess; is_t4 = \"Tesla T4\" in str(subprocess.check_output([\"nvidia-smi\"]))\n", + "except: is_t4 = False\n", + "get_vllm, get_triton = (\"vllm==0.10.1\", \"triton==3.2.0\") if is_t4 else (\"vllm\", \"triton\")\n", + "!uv pip install -qqq --upgrade unsloth {get_vllm} {get_numpy} torchvision bitsandbytes xformers\n", + "!uv pip install -qqq {get_triton}\n", + "!uv pip install \"huggingface_hub>=0.34.0\" \"datasets>=3.4.1,<4.0.\n", + "!uv pip install transformers==4.55.4" + ] }, { "cell_type": "markdown", @@ -193,8 +205,8 @@ "output_type": "stream", "text": [ "Unsloth: Patching Xformers to fix some performance issues.\n", - "\ud83e\udda5 Unsloth: Will patch your computer to enable 2x faster free finetuning.\n", - "\ud83e\udda5 Unsloth Zoo will now patch everything to make training faster!\n", + "🦥 Unsloth: Will patch your computer to enable 2x faster free finetuning.\n", + "🦥 Unsloth Zoo will now patch everything to make training faster!\n", "INFO 03-07 09:37:36 __init__.py:207] Automatically detected platform cuda.\n", "==((====))== Unsloth 2025.3.8: Fast Llama patching. Transformers: 4.48.3. vLLM: 0.7.3.\n", " \\\\ /| Tesla T4. Num GPUs = 1. Max memory: 14.741 GB. Platform: Linux.\n", @@ -346,7 +358,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Capturing CUDA graph shapes: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 23/23 [00:43<00:00, 1.88s/it]" + "Capturing CUDA graph shapes: 100%|██████████| 23/23 [00:43<00:00, 1.88s/it]" ] }, { @@ -456,7 +468,7 @@ }, { "cell_type": "code", - "execution_count": 3, + "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", @@ -684,14 +696,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", @@ -4298,7 +4310,7 @@ "\n", "However, we also need to consider the pants Dani already has. Dani initially had 50 pants.\n", "-------------------- Question:\n", - "Jessica\u2019s class is going to a farm for a field trip. The school will bring all 35 students in the class plus 4 adult chaperones. The farm entrance fee for students costs $5 and $6 for adults. How much will the school pay for the farm entrance in all? \n", + "Jessica’s class is going to a farm for a field trip. The school will bring all 35 students in the class plus 4 adult chaperones. The farm entrance fee for students costs $5 and $6 for adults. How much will the school pay for the farm entrance in all? \n", "Answer:\n", "199 \n", "Response:\n", @@ -4683,7 +4695,7 @@ "\n", "So Emmanuel will get 80 jelly beans.\n", "-------------------- Question:\n", - "Libby has 160 quarters in her piggy bank. She has to pay $35 to replace her sister\u2019s dress that she borrowed and ruined. After replacing the dress, how many quarters will Libby have left? \n", + "Libby has 160 quarters in her piggy bank. She has to pay $35 to replace her sister’s dress that she borrowed and ruined. After replacing the dress, how many quarters will Libby have left? \n", "Answer:\n", "20 \n", "Response:\n", @@ -4699,7 +4711,7 @@ "\n", "Since $5 is equal to 5 / 0.25 = 20 quarters. \n", "\n", - "20 quarters will be left after she pays for the sister\u2019s dress.\n", + "20 quarters will be left after she pays for the sister’s dress.\n", "\n", "20 \n", "Extracted:\n", @@ -4926,7 +4938,7 @@ "Extracted:\n", "21\n", "-------------------- Question:\n", - "Steven is preparing a shipment of boxes to deliver to a customer for his chemical supply business. The products are very delicate and must be carefully packed, so partially filled boxes can\u2019t be shipped. Steven has three trucks that can be sent on each delivery. Each truck can carry a load of no more than 2,000 pounds of cargo. Some of the boxes weigh 10 pounds after being packed, and some of the boxes weigh 40 pounds when packed. Steven\u2019s customer has ordered equal quantities of both the lighter and heavier products. How many boxes of products can Steven ship to his customer in each delivery? \n", + "Steven is preparing a shipment of boxes to deliver to a customer for his chemical supply business. The products are very delicate and must be carefully packed, so partially filled boxes can’t be shipped. Steven has three trucks that can be sent on each delivery. Each truck can carry a load of no more than 2,000 pounds of cargo. Some of the boxes weigh 10 pounds after being packed, and some of the boxes weigh 40 pounds when packed. Steven’s customer has ordered equal quantities of both the lighter and heavier products. How many boxes of products can Steven ship to his customer in each delivery? \n", "Answer:\n", "240 \n", "Response:\n", @@ -4938,12 +4950,12 @@ "\n", "Since each truck can carry a maximum of 2,000 pounds, we can set up the following equation:\n", "\n", - "50x \u2264 2000\n", + "50x ≤ 2000\n", "\n", "Now, we can divide both sides by 50:\n", "\n", - "x \u2264 2000 / 50\n", - "x \u2264 40\n", + "x ≤ 2000 / 50\n", + "x ≤ 40\n", "\n", "Therefore, Steven can ship at most 40 boxes of products per delivery to his customer.\n", "\n", @@ -5161,7 +5173,7 @@ "\n", "Since there are 6 people splitting the cost evenly, the cost per person can be calculated by dividing the total cost by 6:\n", "\n", - "$2340 \u00f7 6 = $390\n", + "$2340 ÷ 6 = $390\n", "$390 \n", "Extracted:\n", "$390\n", @@ -5228,7 +5240,7 @@ "Extracted:\n", "39\n", "-------------------- Question:\n", - "There are 25 different layers of rock in a canyon\u2019s strata. Five of the layers are limestone. Half of the rest are sandstone. Four of the remaining are quartz. Half of the rest are shale. The rest are all different kinds of rock. How many kinds of rock make up only one layer in the canyon\u2019s strata? \n", + "There are 25 different layers of rock in a canyon’s strata. Five of the layers are limestone. Half of the rest are sandstone. Four of the remaining are quartz. Half of the rest are shale. The rest are all different kinds of rock. How many kinds of rock make up only one layer in the canyon’s strata? \n", "Answer:\n", "3 \n", "Response:\n", @@ -5530,7 +5542,7 @@ "\n", "Since the actual value from spending the quarters is $400 and the value from melting is $300, he would get $\\frac{300}{400}$ or $\\frac{3}{4}$, or 0.75 times as much money by melting down the quarters than by spending them.\n", "-------------------- Question:\n", - "Ever since she was a little girl, Sheila has dreamed of traveling the world. To help fund her dream, she bought a large piggy bank in December and started saving. By last week, she had saved $3,000. Pleased with her progress, she has decided to continue saving $276 per month, for 4 years. Today, Sheila\u2019s family secretly added $7,000 into the piggy bank. At the end of 4 years, how much money will be in Sheila\u2019s piggy bank? \n", + "Ever since she was a little girl, Sheila has dreamed of traveling the world. To help fund her dream, she bought a large piggy bank in December and started saving. By last week, she had saved $3,000. Pleased with her progress, she has decided to continue saving $276 per month, for 4 years. Today, Sheila’s family secretly added $7,000 into the piggy bank. At the end of 4 years, how much money will be in Sheila’s piggy bank? \n", "Answer:\n", "23248 \n", "Response:\n", @@ -5548,7 +5560,7 @@ "Extracted:\n", "$23,248.\n", "-------------------- Question:\n", - "Tom\u2019s cat is 8 years old. His rabbit is half the age of his cat. His dog is three times as old as his rabbit. How old is the dog? \n", + "Tom’s cat is 8 years old. His rabbit is half the age of his cat. His dog is three times as old as his rabbit. How old is the dog? \n", "Answer:\n", "12 \n", "Response:\n", @@ -5639,7 +5651,7 @@ "Extracted:\n", "$355\n", "-------------------- Question:\n", - "Haman\u2019s father has an egg business supplying the local market. On a Wednesday morning, his father sends him to go and collect 10 trays of eggs for sale from their store. While loading the eggs into the car, he accidentally drops two trays. He calls his father telling him this, and is told to add 7 more trays for sale. How many eggs were sold that day? \n", + "Haman’s father has an egg business supplying the local market. On a Wednesday morning, his father sends him to go and collect 10 trays of eggs for sale from their store. While loading the eggs into the car, he accidentally drops two trays. He calls his father telling him this, and is told to add 7 more trays for sale. How many eggs were sold that day? \n", "Answer:\n", "540 \n", "Response:\n", @@ -5720,7 +5732,7 @@ "Extracted:\n", "28\n", "-------------------- Question:\n", - "All 20 of Grant\u2019s baby teeth have fallen out and he has a total of $54 from the tooth fairy. Every time he lost a tooth, he put it under his pillow for the tooth fairy, except for one that he dropped on the way home from school and another he swallowed accidentally. The tooth fairy left Grant $20 when he lost his first tooth. How much did the tooth fairy leave him per tooth after his first tooth, assuming equal money exchanged for each tooth thereafter? \n", + "All 20 of Grant’s baby teeth have fallen out and he has a total of $54 from the tooth fairy. Every time he lost a tooth, he put it under his pillow for the tooth fairy, except for one that he dropped on the way home from school and another he swallowed accidentally. The tooth fairy left Grant $20 when he lost his first tooth. How much did the tooth fairy leave him per tooth after his first tooth, assuming equal money exchanged for each tooth thereafter? \n", "Answer:\n", "2 \n", "Response:\n", @@ -6036,7 +6048,7 @@ "Extracted:\n", "3\n", "-------------------- Question:\n", - "After Bella\u2019s ballet performance, she received 2 dozen roses from her parents, and 2 roses from each of her 10 dancer friends. How many roses did Bella receive? \n", + "After Bella’s ballet performance, she received 2 dozen roses from her parents, and 2 roses from each of her 10 dancer friends. How many roses did Bella receive? \n", "Answer:\n", "44 \n", "Response:\n", @@ -6228,14 +6240,14 @@ "Response:\n", "To find the average price per movie, we need to calculate the total cost of the movies and then divide it by the total number of movies.\n", "\n", - "The total cost of the DVDs is 8 \u00d7 $12 = $96.\n", - "The total cost of the Blu-rays is 4 \u00d7 $18 = $72.\n", + "The total cost of the DVDs is 8 × $12 = $96.\n", + "The total cost of the Blu-rays is 4 × $18 = $72.\n", "\n", "The total number of movies Chris bought is 8 + 4 = 12.\n", "\n", "The total cost of all the movies is $96 + $72 = $168.\n", "\n", - "The average price per movie is $168 \u00f7 12 = $14.\n", + "The average price per movie is $168 ÷ 12 = $14.\n", "\n", "14 \n", "Extracted:\n", @@ -6458,8 +6470,8 @@ "Response:\n", "To find the price Janice needs to pay for one juice and one sandwich, we need to first find out the price of one juice and the price of one sandwich.\n", "\n", - "One juice costs $10 \u00f7 5 = $2.\n", - "One sandwich costs $6 \u00f7 2 = $3.\n", + "One juice costs $10 ÷ 5 = $2.\n", + "One sandwich costs $6 ÷ 2 = $3.\n", "\n", "So, in total she would need to pay $2 (for the juice) + $3 (for the sandwich) = $5.\n", "\n", @@ -6502,12 +6514,12 @@ "\n", "He bought 9 cans of tuna for $14.50. To find the cost of one can, we need to divide the total amount he paid by the number of cans.\n", "\n", - "$14.50 \u00f7 9 cans = $1.6111 (rounded to four decimal places).\n", + "$14.50 ÷ 9 cans = $1.6111 (rounded to four decimal places).\n", "\n", "To convert this to cents, we multiply by 100:\n", - "$1.6111 \u00d7 100 = 161.11 cents.\n", + "$1.6111 × 100 = 161.11 cents.\n", "\n", - "Since he used 5 coupons for 25 cents each, he saved a total of 5 \u00d7 25 = 125 cents.\n", + "Since he used 5 coupons for 25 cents each, he saved a total of 5 × 25 = 125 cents.\n", "\n", "Therefore, the actual cost of one can without the coupons would be 161.11 - 1.25 (125 cents) = 161.11 - 1.25 = 160 cents.\n", "\n", @@ -6786,7 +6798,7 @@ "Extracted:\n", "220\n", "-------------------- Question:\n", - "Heath spent his weekend helping at his uncle\u2019s farm planting carrots. He planted 400 rows of carrots with 300 plants in each row. He used a machine to get the planting done and it took him 20 hours. How many carrots did he plant each hour? \n", + "Heath spent his weekend helping at his uncle’s farm planting carrots. He planted 400 rows of carrots with 300 plants in each row. He used a machine to get the planting done and it took him 20 hours. How many carrots did he plant each hour? \n", "Answer:\n", "6,000 \n", "Response:\n", @@ -7040,11 +7052,11 @@ "\n", "Now we multiply both sides of the equation by 3/2.\n", "\n", - "I = 20 \u00d7 3/2 \n", + "I = 20 × 3/2 \n", "I = 30\n", "\n", "Since Dina has twice the number of dolls as Ivy, Dina has 2I dolls.\n", - "Dina has 2 \u00d7 30 = 60 dolls.\n", + "Dina has 2 × 30 = 60 dolls.\n", "\n", "60 \n", "Extracted:\n", @@ -7539,7 +7551,7 @@ "\n", "Since Leonard's age must be an integer, we must have a non-integer age for Jerome and/or Nina. We'll check if their ages can be non-integers.\n", "-------------------- Question:\n", - "Adam's father deposited $2000 in the bank. It receives 8% interest paid throughout the year, and he withdraws the interest as soon as it is deposited. How much will Adam\u2019s father have, including his deposit and the interest received after 2 and a half years? \n", + "Adam's father deposited $2000 in the bank. It receives 8% interest paid throughout the year, and he withdraws the interest as soon as it is deposited. How much will Adam’s father have, including his deposit and the interest received after 2 and a half years? \n", "Answer:\n", "2400 \n", "Response:\n", @@ -7742,7 +7754,7 @@ "Extracted:\n", "25\n", "-------------------- Question:\n", - "Louise is in a toy store. She already has 28 toys worth $10 each in her cart. On her way to the till she adds 20 teddy bears to the cart. If the $580 in Louise\u2019s wallet is exactly enough to pay for all the toys, how much does each teddy bear cost? \n", + "Louise is in a toy store. She already has 28 toys worth $10 each in her cart. On her way to the till she adds 20 teddy bears to the cart. If the $580 in Louise’s wallet is exactly enough to pay for all the toys, how much does each teddy bear cost? \n", "Answer:\n", "15 \n", "Response:\n", @@ -8443,7 +8455,7 @@ "Extracted:\n", "1600\n", "-------------------- Question:\n", - "Julian has 80 Facebook friends. 60% are boys and 40% are girls. Boyd has twice as many friends who are girls and has 100 friends total. What percentage of Boyd\u2019s friends are boys? \n", + "Julian has 80 Facebook friends. 60% are boys and 40% are girls. Boyd has twice as many friends who are girls and has 100 friends total. What percentage of Boyd’s friends are boys? \n", "Answer:\n", "36 \n", "Response:\n", @@ -8497,7 +8509,7 @@ "Extracted:\n", "21600\n", "-------------------- Question:\n", - "Claudia offers art classes to kids and charges $10.00 for her one-hour class. If 20 kids attend Saturday\u2019s class and half that many attend Sunday\u2019s class, how much money does she make? \n", + "Claudia offers art classes to kids and charges $10.00 for her one-hour class. If 20 kids attend Saturday’s class and half that many attend Sunday’s class, how much money does she make? \n", "Answer:\n", "300 \n", "Response:\n", @@ -8729,7 +8741,7 @@ "Extracted:\n", "9\n", "-------------------- Question:\n", - "It\u2019s exam season and Tristan has several exams to prepare for. On Monday, he studies for 4 hours then studies for twice this long on Tuesday. On Wednesday, Thursday, and Friday he studies for 3 hours each day. He wants to study for a total of 25 hours over the week and divides the remaining amount of study time evenly between Saturday and Sunday. How many hours does Tristan spend studying on Saturday? \n", + "It’s exam season and Tristan has several exams to prepare for. On Monday, he studies for 4 hours then studies for twice this long on Tuesday. On Wednesday, Thursday, and Friday he studies for 3 hours each day. He wants to study for a total of 25 hours over the week and divides the remaining amount of study time evenly between Saturday and Sunday. How many hours does Tristan spend studying on Saturday? \n", "Answer:\n", "2 \n", "Response:\n", @@ -8893,7 +8905,7 @@ "Extracted:\n", "48\n", "-------------------- Question:\n", - "Daisy\u2019s milk cow provides 16 cups of milk per day. 75% of the milk is consumed by Daisy\u2019s kids. Daisy uses 50% of the remaining milk to cook with. How much milk is left over? \n", + "Daisy’s milk cow provides 16 cups of milk per day. 75% of the milk is consumed by Daisy’s kids. Daisy uses 50% of the remaining milk to cook with. How much milk is left over? \n", "Answer:\n", "2 \n", "Response:\n", @@ -9046,7 +9058,7 @@ "Extracted:\n", "192\n", "-------------------- Question:\n", - "Erika and her 3 siblings are drawing with chalk outside. Another 3 friends join them and ask if they can help with the drawing. Erika loses 2 pieces of chalk as they are counting and the group realizes there isn't enough chalk for everyone. Erika\u2019s mom brings out another 12 pieces of chalk for them to share and there is now enough chalk for everyone to have 3 pieces each. How many pieces of chalk did Erika and her siblings originally have? \n", + "Erika and her 3 siblings are drawing with chalk outside. Another 3 friends join them and ask if they can help with the drawing. Erika loses 2 pieces of chalk as they are counting and the group realizes there isn't enough chalk for everyone. Erika’s mom brings out another 12 pieces of chalk for them to share and there is now enough chalk for everyone to have 3 pieces each. How many pieces of chalk did Erika and her siblings originally have? \n", "Answer:\n", "11 \n", "Response:\n", @@ -9086,7 +9098,7 @@ "\n", "So, Erika and her siblings originally had 11 pieces of chalk.\n", "-------------------- Question:\n", - "Brian is a real estate agent. He makes a 2% commission on each sale. During June, he sold three houses. The houses sold for $157,000, $499,000, and $125,000. What was Brian\u2019s total commission on these three sales? \n", + "Brian is a real estate agent. He makes a 2% commission on each sale. During June, he sold three houses. The houses sold for $157,000, $499,000, and $125,000. What was Brian’s total commission on these three sales? \n", "Answer:\n", "15,620 \n", "Response:\n", @@ -9193,7 +9205,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Processed prompts: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 1/1 [00:51<00:00, 51.81s/it, est. speed input: 0.75 toks/s, output: 17.01 toks/s]\n" + "Processed prompts: 100%|██████████| 1/1 [00:51<00:00, 51.81s/it, est. speed input: 0.75 toks/s, output: 17.01 toks/s]\n" ] }, { @@ -9202,7 +9214,7 @@ "type": "string" }, "text/plain": [ - "'**Calculating Pi using Python**\\n\\nPi (\u03c0) is a mathematical constant representing the ratio of a circle\\'s circumference to its diameter. Here\\'s a simple and efficient way to calculate an approximation of pi using Python.\\n\\n### Using the Monte Carlo Method\\n\\nThe Monte Carlo method is a computational algorithm that uses random sampling to approximate a value. In this case, we can use it to estimate pi by generating random points within a square and checking if they fall inside a quarter-circle inscribed within it.\\n\\n```python\\nimport random\\nimport math\\n\\ndef estimate_pi(num_samples):\\n \"\"\"\\n Estimate the value of pi using the Monte Carlo method.\\n\\n Args:\\n num_samples (int): The number of random points to generate.\\n\\n Returns:\\n float: An approximation of pi.\\n \"\"\"\\n points_inside_circle = 0\\n\\n for _ in range(num_samples):\\n x, y = random.random(), random.random()\\n distance = x**2 + y**2\\n if distance <= 1:\\n points_inside_circle += 1\\n\\n return (points_inside_circle / num_samples) * 4\\n\\n# Example usage:\\nnum_samples = 1000000\\napprox_pi = estimate_pi(num_samples)\\nprint(f\"Approximation of pi with {num_samples} samples: {approx_pi}\")\\n```\\n\\nThis code generates `num_samples` random points within the square (-1, -1) to (1, 1) and checks if each point falls inside the quarter-circle. The ratio of points inside the circle to the total number of samples multiplied by 4 gives an approximation of pi.\\n\\n### Using the Bailey-Borwein-Plouffe Formula\\n\\nThe Bailey-Borwein-Plouffe (BBP) formula is another way to calculate pi. It\\'s a spigot algorithm that uses a series expansion to generate the digits of pi.\\n\\n```python\\ndef bbp_pi(n):\\n \"\"\"\\n Calculate the Bailey-Borwein-Plouffe formula for pi.\\n\\n Args:\\n n (int): The number of terms to use in the series expansion.\\n\\n Returns:\\n float: An approximation of pi.\\n \"\"\"\\n pi = 0.0\\n for k in range(n):\\n pi += (1 / (16 ** k)) * (\\n 4 / (8 * k + 1) -\\n 2 / (8 * k + 4) -\\n 1 / (8 * k + 5) -\\n 1 / (8 * k + 6)\\n )\\n return pi\\n\\n# Example usage:\\nn = 100\\napprox_pi = bbp_pi(n)\\nprint(f\"Approximation of pi with {n} terms: {approx_pi}\")\\n```\\n\\nThis code calculates an approximation of pi using the BBP formula with `n` terms.\\n\\n### Using the Gauss-Legendre Algorithm\\n\\nThe Gauss-Legendre algorithm is another method for calculating pi. It\\'s a recursive algorithm that uses an iterative process to improve the estimate of pi.\\n\\n```python\\ndef gauss_legendre(n):\\n \"\"\"\\n Calculate pi using the Gauss-Legendre algorithm.\\n\\n Args:\\n n (int): The number of iterations.\\n\\n Returns:\\n float: An approximation of pi.\\n \"\"\"\\n a, b = 1.0, (1 + math.sqrt(5)) / 2.0\\n t = (1 + math.sqrt(5)) / 4.0\\n p = 1.0\\n for _ in range(n):\\n a, b = (a + b) / 2.0, math.sqrt(a * b)\\n t -= p * (a - b)\\n p *= 2.0\\n a = a + t / p\\n b = math.sqrt(b - t / p)\\n return a * b * 16.0\\n\\n# Example usage:\\nn = 100\\napprox_pi = gauss_legendre(n)\\nprint(f\"Approximation of pi with {n} iterations: {approx_pi}\")\\n```\\n\\nThis code calculates an approximation of pi using the Gauss-Legendre algorithm with `n` iterations.\\n\\nRemember that these are simplified examples and may not be the most efficient or accurate methods for calculating pi. For more accurate calculations, you can use specialized libraries or software.'" + "'**Calculating Pi using Python**\\n\\nPi (π) is a mathematical constant representing the ratio of a circle\\'s circumference to its diameter. Here\\'s a simple and efficient way to calculate an approximation of pi using Python.\\n\\n### Using the Monte Carlo Method\\n\\nThe Monte Carlo method is a computational algorithm that uses random sampling to approximate a value. In this case, we can use it to estimate pi by generating random points within a square and checking if they fall inside a quarter-circle inscribed within it.\\n\\n```python\\nimport random\\nimport math\\n\\ndef estimate_pi(num_samples):\\n \"\"\"\\n Estimate the value of pi using the Monte Carlo method.\\n\\n Args:\\n num_samples (int): The number of random points to generate.\\n\\n Returns:\\n float: An approximation of pi.\\n \"\"\"\\n points_inside_circle = 0\\n\\n for _ in range(num_samples):\\n x, y = random.random(), random.random()\\n distance = x**2 + y**2\\n if distance <= 1:\\n points_inside_circle += 1\\n\\n return (points_inside_circle / num_samples) * 4\\n\\n# Example usage:\\nnum_samples = 1000000\\napprox_pi = estimate_pi(num_samples)\\nprint(f\"Approximation of pi with {num_samples} samples: {approx_pi}\")\\n```\\n\\nThis code generates `num_samples` random points within the square (-1, -1) to (1, 1) and checks if each point falls inside the quarter-circle. The ratio of points inside the circle to the total number of samples multiplied by 4 gives an approximation of pi.\\n\\n### Using the Bailey-Borwein-Plouffe Formula\\n\\nThe Bailey-Borwein-Plouffe (BBP) formula is another way to calculate pi. It\\'s a spigot algorithm that uses a series expansion to generate the digits of pi.\\n\\n```python\\ndef bbp_pi(n):\\n \"\"\"\\n Calculate the Bailey-Borwein-Plouffe formula for pi.\\n\\n Args:\\n n (int): The number of terms to use in the series expansion.\\n\\n Returns:\\n float: An approximation of pi.\\n \"\"\"\\n pi = 0.0\\n for k in range(n):\\n pi += (1 / (16 ** k)) * (\\n 4 / (8 * k + 1) -\\n 2 / (8 * k + 4) -\\n 1 / (8 * k + 5) -\\n 1 / (8 * k + 6)\\n )\\n return pi\\n\\n# Example usage:\\nn = 100\\napprox_pi = bbp_pi(n)\\nprint(f\"Approximation of pi with {n} terms: {approx_pi}\")\\n```\\n\\nThis code calculates an approximation of pi using the BBP formula with `n` terms.\\n\\n### Using the Gauss-Legendre Algorithm\\n\\nThe Gauss-Legendre algorithm is another method for calculating pi. It\\'s a recursive algorithm that uses an iterative process to improve the estimate of pi.\\n\\n```python\\ndef gauss_legendre(n):\\n \"\"\"\\n Calculate pi using the Gauss-Legendre algorithm.\\n\\n Args:\\n n (int): The number of iterations.\\n\\n Returns:\\n float: An approximation of pi.\\n \"\"\"\\n a, b = 1.0, (1 + math.sqrt(5)) / 2.0\\n t = (1 + math.sqrt(5)) / 4.0\\n p = 1.0\\n for _ in range(n):\\n a, b = (a + b) / 2.0, math.sqrt(a * b)\\n t -= p * (a - b)\\n p *= 2.0\\n a = a + t / p\\n b = math.sqrt(b - t / p)\\n return a * b * 16.0\\n\\n# Example usage:\\nn = 100\\napprox_pi = gauss_legendre(n)\\nprint(f\"Approximation of pi with {n} iterations: {approx_pi}\")\\n```\\n\\nThis code calculates an approximation of pi using the Gauss-Legendre algorithm with `n` iterations.\\n\\nRemember that these are simplified examples and may not be the most efficient or accurate methods for calculating pi. For more accurate calculations, you can use specialized libraries or software.'" ] }, "execution_count": 6, @@ -9275,7 +9287,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Processed prompts: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 1/1 [00:23<00:00, 23.17s/it, est. speed input: 2.63 toks/s, output: 15.80 toks/s]\n" + "Processed prompts: 100%|██████████| 1/1 [00:23<00:00, 23.17s/it, est. speed input: 2.63 toks/s, output: 15.80 toks/s]\n" ] }, { @@ -9427,7 +9439,7 @@ " \n", " \n", "\n", - " Join Discord if you need help + \u2b50\ufe0f Star us on Github \u2b50\ufe0f\n", + " Join Discord if you need help + ⭐️ Star us on Github ⭐️\n", "\n" ] } @@ -9480,9 +9492,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fda3ed1250d34b12aae8401287322548", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e85f549d21dd4c3293f8239a053b0ad4", - "value": "test-00000-of-00001.parquet:\u2007100%" + "value": "test-00000-of-00001.parquet: 100%" } }, "0722b569fa2e45f9af7d4fb8134e632b": { @@ -9553,9 +9565,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_ef2a701c9c594d4d9ffc8379fa9b5899", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_22c31da7c0564f09bbdb04efdefad11a", - "value": "\u20077473/7473\u2007[00:00<00:00,\u20079130.98\u2007examples/s]" + "value": " 7473/7473 [00:00<00:00, 9130.98 examples/s]" } }, "09fca7e3b90f4b17a2344b67100e26bf": { @@ -9596,9 +9608,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fa155f662d1c40f496daef290ff7a5ff", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_9b2a88f9712d409bbf74bf0c2e14df16", - "value": "\u200717.2M/17.2M\u2007[00:00<00:00,\u2007165MB/s]" + "value": " 17.2M/17.2M [00:00<00:00, 165MB/s]" } }, "0b9dd1c421f14c56986cf30abe0a9455": { @@ -9617,9 +9629,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_73177ee1d64c4f56aa393c45c42e7499", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_98aabe0e4167414d943ff3d7759093ad", - "value": "\u200755.5k/55.5k\u2007[00:00<00:00,\u20073.84MB/s]" + "value": " 55.5k/55.5k [00:00<00:00, 3.84MB/s]" } }, "0bc45e643a9b4a0e8db9f85180c108b2": { @@ -9947,9 +9959,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_d4eeb17c3a124c919058c0ffe72a60b6", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_8b6bd37f64454464b44691ae3bef0b9f", - "value": "Loading\u2007safetensors\u2007checkpoint\u2007shards:\u2007100%\u2007Completed\u2007|\u20071/1\u2007[00:27<00:00,\u200727.33s/it]\n" + "value": "Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:27<00:00, 27.33s/it]\n" } }, "18d2fd154ffc4671b76f00760eb35571": { @@ -9983,9 +9995,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_594a583be494419c845fd873904d4d6d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_ba21e0c745714031aae12b1665847e54", - "value": "tokenizer.json:\u2007100%" + "value": "tokenizer.json: 100%" } }, "1a090b75a50b459a8d2f822efc692dfa": { @@ -10184,9 +10196,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1a090b75a50b459a8d2f822efc692dfa", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_0da6c9fa4eeb4b4d832f8bbc614bc8d9", - "value": "\u20077.94k/7.94k\u2007[00:00<00:00,\u2007557kB/s]" + "value": " 7.94k/7.94k [00:00<00:00, 557kB/s]" } }, "22c31da7c0564f09bbdb04efdefad11a": { @@ -10339,9 +10351,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_364d9414fd5f4280be4ad621874e15cb", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_7b3ad76804bd47a68fba0bf9f967cb04", - "value": "\u20075.96G/5.96G\u2007[00:45<00:00,\u2007545MB/s]" + "value": " 5.96G/5.96G [00:45<00:00, 545MB/s]" } }, "2aade13ff2ca4250ae02b7a12b1ef16d": { @@ -10875,9 +10887,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_79da5ea5195c4c63b0ec1ee354511a87", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_a5f3704359df48e4b4c5b418a5f72021", - "value": "\u2007454/454\u2007[00:00<00:00,\u200724.6kB/s]" + "value": " 454/454 [00:00<00:00, 24.6kB/s]" } }, "397a49fc365b4c3a95db709b71fc32e3": { @@ -10948,9 +10960,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_0ea62e14a2d44b8ca7c6f15bca8e1a2b", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_27e78ef30ff64be5af79b6431816f569", - "value": "\u2007239/239\u2007[00:00<00:00,\u200723.8kB/s]" + "value": " 239/239 [00:00<00:00, 23.8kB/s]" } }, "436282df7bcc45c4b951de60f2e4ec53": { @@ -11021,9 +11033,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_269f762b5af041eab750226febe70e07", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_f38c687b78224c8aa2227914998b4627", - "value": "train-00000-of-00001.parquet:\u2007100%" + "value": "train-00000-of-00001.parquet: 100%" } }, "48403d21996b4ebfa931cf260af6c283": { @@ -11146,9 +11158,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fe9dbcf4e85646cb8c684d82252b93e2", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_d36f0dda703c47cfae6f544e7e655806", - "value": "\u200755.5k/55.5k\u2007[00:00<00:00,\u20074.95MB/s]" + "value": " 55.5k/55.5k [00:00<00:00, 4.95MB/s]" } }, "4c69d53cf33c4865af8feead04b810ad": { @@ -11437,9 +11449,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_5e9d2ad532a8447a904688629a91f171", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_2df38622a5ab46979ee53efeb8c31b91", - "value": "README.md:\u2007100%" + "value": "README.md: 100%" } }, "5de480828775465bb76b674ecd713bd0": { @@ -11562,9 +11574,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a67ca6e6c91244978b6b468257ffefd7", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_781f0652012b424aa43cd39028e3a99a", - "value": "generation_config.json:\u2007100%" + "value": "generation_config.json: 100%" } }, "638c59bf00ba4536ad157975faa26984": { @@ -11661,9 +11673,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_81f43f522faa4dcf9e557793a2027e90", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_ae46c8f17c0644fdbd1aa6d412c322b7", - "value": "Generating\u2007train\u2007split:\u2007100%" + "value": "Generating train split: 100%" } }, "6be20c9739de44ffb7ef2465ad17424d": { @@ -11721,9 +11733,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_bc92bf99bcf345b59699eb8f0423f7cd", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_f589e115f63d4fdb95816dccb34c3e3e", - "value": "tokenizer_config.json:\u2007100%" + "value": "tokenizer_config.json: 100%" } }, "70780d08fb00445da32b28b40d33412f": { @@ -12048,9 +12060,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_ff032a23470845ba95a12227baf0e0ce", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_18d2fd154ffc4671b76f00760eb35571", - "value": "\u200717.2M/17.2M\u2007[00:00<00:00,\u2007103MB/s]" + "value": " 17.2M/17.2M [00:00<00:00, 103MB/s]" } }, "7b3ad76804bd47a68fba0bf9f967cb04": { @@ -12084,7 +12096,7 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_7083c21ed3dd4ed6b442afdf4a18a05d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_6e4ce08c86f24beeb63d38d436143dc9", "value": "" } @@ -12181,9 +12193,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_cbe9c4e2f8bc49dfb1d8a4cddc36d724", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_f33aef89be7641a694fb11e3101c7d85", - "value": "\u2007454/454\u2007[00:00<00:00,\u200739.2kB/s]" + "value": " 454/454 [00:00<00:00, 39.2kB/s]" } }, "813e8e40df6447c18e47414ab38f69a2": { @@ -12202,9 +12214,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_115a8f3c523c4cd3aaa472af87fe0e4a", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_26cde13f97db4ae4a0900fb5184a51a2", - "value": "special_tokens_map.json:\u2007100%" + "value": "special_tokens_map.json: 100%" } }, "81f43f522faa4dcf9e557793a2027e90": { @@ -12327,9 +12339,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_32fe42f8a11d4131a3a529b554f028fe", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_39c33d513fb74712be8fcf73093a0c9e", - "value": "\u20071319/1319\u2007[00:00<00:00,\u200728401.73\u2007examples/s]" + "value": " 1319/1319 [00:00<00:00, 28401.73 examples/s]" } }, "8b6bd37f64454464b44691ae3bef0b9f": { @@ -12546,9 +12558,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_d89eae07da0e4344909900e24e3d0d09", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_b1e00fc06b664fc59d4921388f187269", - "value": "Map:\u2007100%" + "value": "Map: 100%" } }, "93f708b722194750a9f66a2597987033": { @@ -12805,9 +12817,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_b095d3eac38847c0b087ffcc5876e0b6", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_9ebd8c869bc144cf8142375e7d7f04dd", - "value": "tokenizer.json:\u2007100%" + "value": "tokenizer.json: 100%" } }, "9ec592e894b945cea7c2a756a96eb959": { @@ -12879,9 +12891,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fac3506146244766b886c715ea57ec7b", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_d068e96c067e43ffa1777a666b0df147", - "value": "tokenizer_config.json:\u2007100%" + "value": "tokenizer_config.json: 100%" } }, "a5f3704359df48e4b4c5b418a5f72021": { @@ -13071,9 +13083,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_436282df7bcc45c4b951de60f2e4ec53", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_911d088df0894243b69fd4f69c46e7f3", - "value": "\u2007419k/419k\u2007[00:00<00:00,\u20078.72MB/s]" + "value": " 419k/419k [00:00<00:00, 8.72MB/s]" } }, "ae46c8f17c0644fdbd1aa6d412c322b7": { @@ -13385,9 +13397,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4c69d53cf33c4865af8feead04b810ad", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_9dbc67661c614d8a82d0790d9b694e6d", - "value": "Loading\u2007safetensors\u2007checkpoint\u2007shards:\u2007100%\u2007Completed\u2007|\u20071/1\u2007[00:14<00:00,\u200714.38s/it]\n" + "value": "Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:14<00:00, 14.38s/it]\n" } }, "c3ec8a75007846519d40f6726d2a01ee": { @@ -13813,9 +13825,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_2aade13ff2ca4250ae02b7a12b1ef16d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_23531535f2fc4b99a5ecd1b63fb93d80", - "value": "\u20072.31M/2.31M\u2007[00:00<00:00,\u200746.9MB/s]" + "value": " 2.31M/2.31M [00:00<00:00, 46.9MB/s]" } }, "d4eeb17c3a124c919058c0ffe72a60b6": { @@ -13938,9 +13950,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_2df204a0982b466eb3c25bf80645b6c1", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_9ec592e894b945cea7c2a756a96eb959", - "value": "special_tokens_map.json:\u2007100%" + "value": "special_tokens_map.json: 100%" } }, "d7e3f2c2441743089ee7ed10eb0319dd": { @@ -13959,9 +13971,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_2c83de81fc91497db859d1be6860d6f4", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_397a49fc365b4c3a95db709b71fc32e3", - "value": "model.safetensors:\u2007100%" + "value": "model.safetensors: 100%" } }, "d89eae07da0e4344909900e24e3d0d09": { @@ -14116,9 +14128,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_cc6097b25ec74decb51a397ccd9fcf27", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_d2af3c19e6ec46a3915f6badf7062b7b", - "value": "\u20077473/7473\u2007[00:00<00:00,\u20072457.10\u2007examples/s]" + "value": " 7473/7473 [00:00<00:00, 2457.10 examples/s]" } }, "e1e8c601a2bb4c0b8f0cdf3bf0dd0b2e": { @@ -14243,7 +14255,7 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fe4ecfe3522c432b964608f137cfb936", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e1e8c601a2bb4c0b8f0cdf3bf0dd0b2e", "value": "" } @@ -14551,9 +14563,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_76dc9524ee764741a1615de6bc43492e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_78201e7953b647ee9656fc7e8f4abbcb", - "value": "Generating\u2007test\u2007split:\u2007100%" + "value": "Generating test split: 100%" } }, "fa155f662d1c40f496daef290ff7a5ff": { @@ -14926,4 +14938,4 @@ }, "nbformat": 4, "nbformat_minor": 0 -} \ No newline at end of file +} diff --git a/nb/Kaggle-Mistral_v0.3_(7B)-GRPO.ipynb b/nb/Kaggle-Mistral_v0.3_(7B)-GRPO.ipynb index 4f32f0b..c27129c 100644 --- a/nb/Kaggle-Mistral_v0.3_(7B)-GRPO.ipynb +++ b/nb/Kaggle-Mistral_v0.3_(7B)-GRPO.ipynb @@ -536,14 +536,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", diff --git a/nb/Kaggle-Phi_4_(14B)-GRPO.ipynb b/nb/Kaggle-Phi_4_(14B)-GRPO.ipynb index 7276002..3222b0f 100644 --- a/nb/Kaggle-Phi_4_(14B)-GRPO.ipynb +++ b/nb/Kaggle-Phi_4_(14B)-GRPO.ipynb @@ -8,7 +8,7 @@ "
\n", "\n", "\n", - " Join Discord if you need help + \u2b50 Star us on Github \u2b50\n", + " Join Discord if you need help + ⭐ Star us on Github ⭐\n", "
\n", "\n", "To install Unsloth on your own computer, follow the installation instructions on our Github page [here](https://docs.unsloth.ai/get-started/installing-+-updating).\n", @@ -48,7 +48,19 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\n!pip install --upgrade -qqq uv\ntry: import numpy; get_numpy = f\"numpy=={numpy.__version__}\"\nexcept: get_numpy = \"numpy\"\ntry: import subprocess; is_t4 = \"Tesla T4\" in str(subprocess.check_output([\"nvidia-smi\"]))\nexcept: is_t4 = False\nget_vllm, get_triton = (\"vllm==0.10.1\", \"triton==3.2.0\") if is_t4 else (\"vllm\", \"triton\")\n!uv pip install -qqq --upgrade unsloth {get_vllm} {get_numpy} torchvision bitsandbytes xformers\n!uv pip install -qqq {get_triton}\n!uv pip install \"huggingface_hub>=0.34.0\" \"datasets>=3.4.1,<4.0.\n!uv pip install transformers==4.55.4" + "source": [ + "%%capture\n", + "!pip install --upgrade -qqq uv\n", + "try: import numpy; get_numpy = f\"numpy=={numpy.__version__}\"\n", + "except: get_numpy = \"numpy\"\n", + "try: import subprocess; is_t4 = \"Tesla T4\" in str(subprocess.check_output([\"nvidia-smi\"]))\n", + "except: is_t4 = False\n", + "get_vllm, get_triton = (\"vllm==0.10.1\", \"triton==3.2.0\") if is_t4 else (\"vllm\", \"triton\")\n", + "!uv pip install -qqq --upgrade unsloth {get_vllm} {get_numpy} torchvision bitsandbytes xformers\n", + "!uv pip install -qqq {get_triton}\n", + "!uv pip install \"huggingface_hub>=0.34.0\" \"datasets>=3.4.1,<4.0.\n", + "!uv pip install transformers==4.55.4" + ] }, { "cell_type": "markdown", @@ -175,7 +187,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Capturing CUDA graph shapes: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 19/19 [00:50<00:00, 2.67s/it]" + "Capturing CUDA graph shapes: 100%|██████████| 19/19 [00:50<00:00, 2.67s/it]" ] }, { @@ -308,14 +320,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", @@ -1491,11 +1503,11 @@ "\n", "1. **Calculate the cost of the books:**\n", " - Lynne bought 7 books about cats and 2 books about the solar system. This is a total of 7 + 2 = 9 books.\n", - " - Each book costs $7. So, the total cost for books is 9 books \u00d7 $7/book = $63.\n", + " - Each book costs $7. So, the total cost for books is 9 books × $7/book = $63.\n", "\n", "2. **Calculate the cost of the magazines:**\n", " - Lynne bought 3 magazines.\n", - " - Each magazine costs $4. So, the total cost for magazines is 3 magazines \u00d7 $4/magazine = $12.\n", + " - Each magazine costs $4. So, the total cost for magazines is 3 magazines × $4/magazine = $12.\n", "\n", "3. **Calculate the total amount spent:**\n", " - Add the total cost of the books and the magazines: $63 (books) + $12 (magazines) = $75.\n", @@ -1507,11 +1519,11 @@ "\n", "1. **Calculate the cost of the books:**\n", " - Lynne bought 7 books about cats and 2 books about the solar system. This is a total of 7 + 2 = 9 books.\n", - " - Each book costs $7. So, the total cost for books is 9 books \u00d7 $7/book = $63.\n", + " - Each book costs $7. So, the total cost for books is 9 books × $7/book = $63.\n", "\n", "2. **Calculate the cost of the magazines:**\n", " - Lynne bought 3 magazines.\n", - " - Each magazine costs $4. So, the total cost for magazines is 3 magazines \u00d7 $4/magazine = $12.\n", + " - Each magazine costs $4. So, the total cost for magazines is 3 magazines × $4/magazine = $12.\n", "\n", "3. **Calculate the total amount spent:**\n", " - Add the total cost of the books and the magazines: $63 (books) + $12 (magazines) = $75.\n", @@ -1649,7 +1661,7 @@ "Extracted:\n", "21\n", "-------------------- Question:\n", - "There are 84 people waiting in line to ride a roller coaster at an amusement park.\u00a0 The roller coaster has 7 cars, and each car seats 2 people.\u00a0 How many times will the ride operator have to run the roller coaster to give everyone in line a turn? \n", + "There are 84 people waiting in line to ride a roller coaster at an amusement park.  The roller coaster has 7 cars, and each car seats 2 people.  How many times will the ride operator have to run the roller coaster to give everyone in line a turn? \n", "Answer:\n", "6 \n", "Response:\n", @@ -1731,17 +1743,17 @@ "1. **Necklaces:**\n", " - Price per necklace: $25.00\n", " - Number sold: 5\n", - " - Total from necklaces = 5 necklaces \u00d7 $25.00/necklace = $125.00\n", + " - Total from necklaces = 5 necklaces × $25.00/necklace = $125.00\n", "\n", "2. **Bracelets:**\n", " - Price per bracelet: $15.00\n", " - Number sold: 10\n", - " - Total from bracelets = 10 bracelets \u00d7 $15.00/bracelet = $150.00\n", + " - Total from bracelets = 10 bracelets × $15.00/bracelet = $150.00\n", "\n", "3. **Earrings:**\n", " - Price per pair of earrings: $10.00\n", " - Number sold: 20\n", - " - Total from earrings = 20 pairs \u00d7 $10.00/pair = $200.00\n", + " - Total from earrings = 20 pairs × $10.00/pair = $200.00\n", "\n", "4. **Complete Jewelry Ensembles:**\n", " - Price per ensemble: \n", @@ -1752,17 +1764,17 @@ "1. **Necklaces:**\n", " - Price per necklace: $25.00\n", " - Number sold: 5\n", - " - Total from necklaces = 5 necklaces \u00d7 $25.00/necklace = $125.00\n", + " - Total from necklaces = 5 necklaces × $25.00/necklace = $125.00\n", "\n", "2. **Bracelets:**\n", " - Price per bracelet: $15.00\n", " - Number sold: 10\n", - " - Total from bracelets = 10 bracelets \u00d7 $15.00/bracelet = $150.00\n", + " - Total from bracelets = 10 bracelets × $15.00/bracelet = $150.00\n", "\n", "3. **Earrings:**\n", " - Price per pair of earrings: $10.00\n", " - Number sold: 20\n", - " - Total from earrings = 20 pairs \u00d7 $10.00/pair = $200.00\n", + " - Total from earrings = 20 pairs × $10.00/pair = $200.00\n", "\n", "4. **Complete Jewelry Ensembles:**\n", " - Price per ensemble:\n", @@ -1831,7 +1843,7 @@ " Convert 0.5 to a fraction with a common denominator: \n", " \\( 0.5 = \\frac{3}{6} = \\frac{4}{6} = \\frac{2}{3}\n", "-------------------- Question:\n", - "At Hank\u2019s cafe, he sells big stack pancakes which have 5 pancakes and short stack pancakes which have 3 pancakes. If 9 customers order the short stack and 6 customers order the big stack, how many pancakes does Hank need to make? \n", + "At Hank’s cafe, he sells big stack pancakes which have 5 pancakes and short stack pancakes which have 3 pancakes. If 9 customers order the short stack and 6 customers order the big stack, how many pancakes does Hank need to make? \n", "Answer:\n", "57 \n", "Response:\n", @@ -2094,8 +2106,8 @@ "To determine how many candies are in each packet, we first need to find out how many candies Bobby eats in 3 weeks.\n", "\n", "Bobby's candy consumption breakdown for one week is as follows:\n", - "- Monday to Friday: 2 candies per day \u00d7 5 days = 10 candies\n", - "- Saturday and Sunday: 1 candy per day \u00d7 2 days = 2 candies\n", + "- Monday to Friday: 2 candies per day × 5 days = 10 candies\n", + "- Saturday and Sunday: 1 candy per day × 2 days = 2 candies\n", "\n", "Therefore, in one week, Bobby eats:\n", "\\[ 10 \\text{ candies} + 2 \\text{ candies} = 12 \\text{ candies} \\]\n", @@ -2109,8 +2121,8 @@ "To determine how many candies are in each packet, we first need to find out how many candies Bobby eats in 3 weeks.\n", "\n", "Bobby's candy consumption breakdown for one week is as follows:\n", - "- Monday to Friday: 2 candies per day \u00d7 5 days = 10 candies\n", - "- Saturday and Sunday: 1 candy per day \u00d7 2 days = 2 candies\n", + "- Monday to Friday: 2 candies per day × 5 days = 10 candies\n", + "- Saturday and Sunday: 1 candy per day × 2 days = 2 candies\n", "\n", "Therefore, in one week, Bobby eats:\n", "\\[ 10 \\text{ candies} + 2 \\text{ candies} = 12 \\text{ candies} \\]\n", @@ -2702,7 +2714,7 @@ "Extracted:\n", "20 counselors\n", "-------------------- Question:\n", - "After Bella\u2019s ballet performance, she received 2 dozen roses from her parents, and 2 roses from each of her 10 dancer friends. How many roses did Bella receive? \n", + "After Bella’s ballet performance, she received 2 dozen roses from her parents, and 2 roses from each of her 10 dancer friends. How many roses did Bella receive? \n", "Answer:\n", "44 \n", "Response:\n", @@ -3730,16 +3742,16 @@ " \\[\n", " 2M - 1 = 15 \\quad \\text{(Multiplying both sides by 3)}\n", "-------------------- Question:\n", - "Clyde and Grace are building block towers. Grace\u2019s tower is 8 times the size of Clyde\u2019s at 40 inches tall. How many inches taller is Grace\u2019s tower than Clyde\u2019s? \n", + "Clyde and Grace are building block towers. Grace’s tower is 8 times the size of Clyde’s at 40 inches tall. How many inches taller is Grace’s tower than Clyde’s? \n", "Answer:\n", "35 \n", "Response:\n", "\n", "To solve this problem, we need to determine the height of Clyde's tower and compare it to the height of Grace's tower.\n", "\n", - "1. We know that Grace\u2019s tower is 8 times the size of Clyde\u2019s tower. Let \\( x \\) be the height of Clyde's tower. Therefore, Grace's tower is \\( 8x \\).\n", + "1. We know that Grace’s tower is 8 times the size of Clyde’s tower. Let \\( x \\) be the height of Clyde's tower. Therefore, Grace's tower is \\( 8x \\).\n", "\n", - "2. We are given that Grace\u2019s tower is 40 inches tall. So, we can write the equation:\n", + "2. We are given that Grace’s tower is 40 inches tall. So, we can write the equation:\n", " \\[\n", " 8x = 40\n", " \\]\n", @@ -3751,14 +3763,14 @@ "\n", "4. Clyde's tower is 5 inches tall.\n", "\n", - "5. To find how many inches taller Grace\u2019s tower is compared to Clyde's, we subtract the height of Clyde\u2019s tower \n", + "5. To find how many inches taller Grace’s tower is compared to Clyde's, we subtract the height of Clyde’s tower \n", "Extracted:\n", "\n", "To solve this problem, we need to determine the height of Clyde's tower and compare it to the height of Grace's tower.\n", "\n", - "1. We know that Grace\u2019s tower is 8 times the size of Clyde\u2019s tower. Let \\( x \\) be the height of Clyde's tower. Therefore, Grace's tower is \\( 8x \\).\n", + "1. We know that Grace’s tower is 8 times the size of Clyde’s tower. Let \\( x \\) be the height of Clyde's tower. Therefore, Grace's tower is \\( 8x \\).\n", "\n", - "2. We are given that Grace\u2019s tower is 40 inches tall. So, we can write the equation:\n", + "2. We are given that Grace’s tower is 40 inches tall. So, we can write the equation:\n", " \\[\n", " 8x = 40\n", " \\]\n", @@ -3770,7 +3782,7 @@ "\n", "4. Clyde's tower is 5 inches tall.\n", "\n", - "5. To find how many inches taller Grace\u2019s tower is compared to Clyde's, we subtract the height of Clyde\u2019s tower\n", + "5. To find how many inches taller Grace’s tower is compared to Clyde's, we subtract the height of Clyde’s tower\n", "-------------------- Question:\n", "Andrew's father buys a package of 100 masks. Andrew lives with his 2 parents and 2 siblings. All members of Andrew's family change masks every 4 days. How many days will it take to finish the pack of masks? \n", "Answer:\n", @@ -3786,7 +3798,7 @@ "2. Each member of the family changes their mask every 4 days. This implies that each person uses 1 mask every 4 days.\n", "\n", "3. Therefore, in a 4-day period, all 5 family members together would use:\n", - " - 5 people \u00d7 1 mask per person = 5 masks.\n", + " - 5 people × 1 mask per person = 5 masks.\n", "\n", "4. Now, we want to know how many total days it will take for the package of 100 masks to be used up. Let's calculate how many 4-day periods it takes to use 100 masks:\n", " - Number of 4-day \n", @@ -3801,7 +3813,7 @@ "2. Each member of the family changes their mask every 4 days. This implies that each person uses 1 mask every 4 days.\n", "\n", "3. Therefore, in a 4-day period, all 5 family members together would use:\n", - " - 5 people \u00d7 1 mask per person = 5 masks.\n", + " - 5 people × 1 mask per person = 5 masks.\n", "\n", "4. Now, we want to know how many total days it will take for the package of 100 masks to be used up. Let's calculate how many 4-day periods it takes to use 100 masks:\n", " - Number of 4-day\n", @@ -4040,7 +4052,7 @@ "\n", "Given that \\( x = 210 \\), and we know that after cutting these \\( x \\) bananas\n", "-------------------- Question:\n", - "At Theo\u2019s cafe, he makes 3 egg and 4 egg omelettes. His cafe is open from 7:00 a.m. to 11:00 a.m. In the first hour, 5 customers order the 3 egg omelettes. In the second hour, 7 customers order the 4 egg omelettes. In the third hour, 3 customers order the 3 egg omelettes. In the last hour, 8 customers order the 4 egg omelettes. How many eggs does Theo need to make all the omelettes? \n", + "At Theo’s cafe, he makes 3 egg and 4 egg omelettes. His cafe is open from 7:00 a.m. to 11:00 a.m. In the first hour, 5 customers order the 3 egg omelettes. In the second hour, 7 customers order the 4 egg omelettes. In the third hour, 3 customers order the 3 egg omelettes. In the last hour, 8 customers order the 4 egg omelettes. How many eggs does Theo need to make all the omelettes? \n", "Answer:\n", "84 \n", "Response:\n", @@ -4429,7 +4441,7 @@ "\\]\n", "\n", "Now, calculate the number of slices the waiter receives:\n", - "- Waiter's slices = 8 parts \u00d7 6 slices per part = 48 slices\n", + "- Waiter's slices = 8 parts × 6 slices per part = 48 slices\n", "\n", "Finally, we need to find \"twenty less than the number of slices the waiter ate\":\n", "- Slices the waiter ate = 48\n", @@ -4449,7 +4461,7 @@ "\\]\n", "\n", "Now, calculate the number of slices the waiter receives:\n", - "- Waiter's slices = 8 parts \u00d7 6 slices per part = 48 slices\n", + "- Waiter's slices = 8 parts × 6 slices per part = 48 slices\n", "\n", "Finally, we need to find \"twenty less than the number of slices the waiter ate\":\n", "- Slices the waiter ate = 48\n", @@ -4966,7 +4978,7 @@ "2. **Calculate the weekly equivalent of the monthly housing benefit reduction:**\n", " - The monthly reduction in his housing benefit is $60.\n", " - To find the equivalent weekly reduction, divide the monthly reduction by the average number of weeks in a month (approximately 4.33 weeks per month).\n", - " - Weekly reduction in housing benefit = $60/month \u00f7 4.33 weeks/month \u2248 $13.86/week.\n", + " - Weekly reduction in housing benefit = $60/month ÷ 4.33 weeks/month ≈ $13.86/week.\n", "\n", "3. **Calculate the net weekly effect:**\n", " - Net weekly \n", @@ -4982,7 +4994,7 @@ "2. **Calculate the weekly equivalent of the monthly housing benefit reduction:**\n", " - The monthly reduction in his housing benefit is $60.\n", " - To find the equivalent weekly reduction, divide the monthly reduction by the average number of weeks in a month (approximately 4.33 weeks per month).\n", - " - Weekly reduction in housing benefit = $60/month \u00f7 4.33 weeks/month \u2248 $13.86/week.\n", + " - Weekly reduction in housing benefit = $60/month ÷ 4.33 weeks/month ≈ $13.86/week.\n", "\n", "3. **Calculate the net weekly effect:**\n", " - Net weekly\n", @@ -5116,7 +5128,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Processed prompts: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 1/1 [00:16<00:00, 16.59s/it, est. speed input: 1.27 toks/s, output: 9.89 toks/s]\n" + "Processed prompts: 100%|██████████| 1/1 [00:16<00:00, 16.59s/it, est. speed input: 1.27 toks/s, output: 9.89 toks/s]\n" ] }, { @@ -5198,7 +5210,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Processed prompts: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 1/1 [00:27<00:00, 27.72s/it, est. speed input: 1.70 toks/s, output: 10.03 toks/s]\n" + "Processed prompts: 100%|██████████| 1/1 [00:27<00:00, 27.72s/it, est. speed input: 1.70 toks/s, output: 10.03 toks/s]\n" ] }, { @@ -5396,7 +5408,7 @@ " \n", " \n", "\n", - " Join Discord if you need help + \u2b50\ufe0f Star us on Github \u2b50\ufe0f\n", + " Join Discord if you need help + ⭐️ Star us on Github ⭐️\n", "\n" ] } @@ -5678,7 +5690,7 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_e43ad27d5d304d1ebf9b374016409a97", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_51948945111f437c9ed6ccab22072dd3", "value": "" } @@ -5699,7 +5711,7 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4c9248100f89400d9e1407dbb168d5d6", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_96cea0d773c8426b8be72dd7f72e5a82", "value": "" } @@ -5949,9 +5961,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_9eed940f3815428583b4ddefc1a81469", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_0b9230e976b34a9ea85978cf22857012", - "value": "Loading\u2007safetensors\u2007checkpoint\u2007shards:\u2007100%\u2007Completed\u2007|\u20072/2\u2007[00:51<00:00,\u200725.46s/it]\n" + "value": "Loading safetensors checkpoint shards: 100% Completed | 2/2 [00:51<00:00, 25.46s/it]\n" } }, "d034c840e7f74177a7b07a188d666b8d": { @@ -5970,9 +5982,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_129dc789722b43439574390bba63b36a", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_8991360910ef417db03499f76f5fe323", - "value": "Loading\u2007safetensors\u2007checkpoint\u2007shards:\u2007100%\u2007Completed\u2007|\u20072/2\u2007[00:47<00:00,\u200723.63s/it]\n" + "value": "Loading safetensors checkpoint shards: 100% Completed | 2/2 [00:47<00:00, 23.63s/it]\n" } }, "d6d5a7d96a034247b38d25d8a9cc979c": { @@ -6107,4 +6119,4 @@ }, "nbformat": 4, "nbformat_minor": 0 -} \ No newline at end of file +} diff --git a/nb/Kaggle-Qwen2.5_(3B)-GRPO.ipynb b/nb/Kaggle-Qwen2.5_(3B)-GRPO.ipynb index acfe3ca..392972e 100644 --- a/nb/Kaggle-Qwen2.5_(3B)-GRPO.ipynb +++ b/nb/Kaggle-Qwen2.5_(3B)-GRPO.ipynb @@ -8,7 +8,7 @@ "
\n", "\n", "\n", - " Join Discord if you need help + \u2b50 Star us on Github \u2b50\n", + " Join Discord if you need help + ⭐ Star us on Github ⭐\n", "
\n", "\n", "To install Unsloth on your own computer, follow the installation instructions on our Github page [here](https://docs.unsloth.ai/get-started/installing-+-updating).\n", @@ -48,7 +48,19 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\n!pip install --upgrade -qqq uv\ntry: import numpy; get_numpy = f\"numpy=={numpy.__version__}\"\nexcept: get_numpy = \"numpy\"\ntry: import subprocess; is_t4 = \"Tesla T4\" in str(subprocess.check_output([\"nvidia-smi\"]))\nexcept: is_t4 = False\nget_vllm, get_triton = (\"vllm==0.10.1\", \"triton==3.2.0\") if is_t4 else (\"vllm\", \"triton\")\n!uv pip install -qqq --upgrade unsloth {get_vllm} {get_numpy} torchvision bitsandbytes xformers\n!uv pip install -qqq {get_triton}\n!uv pip install \"huggingface_hub>=0.34.0\" \"datasets>=3.4.1,<4.0.\n!uv pip install transformers==4.55.4" + "source": [ + "%%capture\n", + "!pip install --upgrade -qqq uv\n", + "try: import numpy; get_numpy = f\"numpy=={numpy.__version__}\"\n", + "except: get_numpy = \"numpy\"\n", + "try: import subprocess; is_t4 = \"Tesla T4\" in str(subprocess.check_output([\"nvidia-smi\"]))\n", + "except: is_t4 = False\n", + "get_vllm, get_triton = (\"vllm==0.10.1\", \"triton==3.2.0\") if is_t4 else (\"vllm\", \"triton\")\n", + "!uv pip install -qqq --upgrade unsloth {get_vllm} {get_numpy} torchvision bitsandbytes xformers\n", + "!uv pip install -qqq {get_triton}\n", + "!uv pip install \"huggingface_hub>=0.34.0\" \"datasets>=3.4.1,<4.0.\n", + "!uv pip install transformers==4.55.4" + ] }, { "cell_type": "markdown", @@ -445,7 +457,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Capturing CUDA graph shapes: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 27/27 [00:44<00:00, 1.66s/it]" + "Capturing CUDA graph shapes: 100%|██████████| 27/27 [00:44<00:00, 1.66s/it]" ] }, { @@ -825,14 +837,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", @@ -965,7 +977,7 @@ "\n", "Emily currently has 9 assignments with a total score of 92 points, and the final assignment she gets is 90 points. Therefore, her current total is:\n", "\\[ 9 \\times 92 + 90 = 828 + 90 = 918 \\]\n", - "She needs to find the minimum score to beat Ahmed, and her target would be greater than Adam\u2019s 813 points from his 9 assignments, so we calculate his total:\n", + "She needs to find the minimum score to beat Ahmed, and her target would be greater than Adam’s 813 points from his 9 assignments, so we calculate his total:\n", "\\[ 9 \\times 91 = 819 \\]\n", "To beat Emily, Ahmed needs a score such that his average score from all 10 \n", "Extracted:\n", @@ -974,7 +986,7 @@ "\n", "Emily currently has 9 assignments with a total score of 92 points, and the final assignment she gets is 90 points. Therefore, her current total is:\n", "\\[ 9 \\times 92 + 90 = 828 + 90 = 918 \\]\n", - "She needs to find the minimum score to beat Ahmed, and her target would be greater than Adam\u2019s 813 points from his 9 assignments, so we calculate his total:\n", + "She needs to find the minimum score to beat Ahmed, and her target would be greater than Adam’s 813 points from his 9 assignments, so we calculate his total:\n", "\\[ 9 \\times 91 = 819 \\]\n", "To beat Emily, Ahmed needs a score such that his average score from all 10\n" ] @@ -3364,7 +3376,7 @@ "To find \\( x \\), we solve:\n", "\\[ 5 = \\frac{1}{\n", "-------------------- Question:\n", - "Clyde and Grace are building block towers. Grace\u2019s tower is 8 times the size of Clyde\u2019s at 40 inches tall. How many inches taller is Grace\u2019s tower than Clyde\u2019s? \n", + "Clyde and Grace are building block towers. Grace’s tower is 8 times the size of Clyde’s at 40 inches tall. How many inches taller is Grace’s tower than Clyde’s? \n", "Answer:\n", "35 \n", "Response:\n", @@ -3605,7 +3617,7 @@ "\n", "\\[ \\text{Initial Number of Bananas} -\n", "-------------------- Question:\n", - "At Theo\u2019s cafe, he makes 3 egg and 4 egg omelettes. His cafe is open from 7:00 a.m. to 11:00 a.m. In the first hour, 5 customers order the 3 egg omelettes. In the second hour, 7 customers order the 4 egg omelettes. In the third hour, 3 customers order the 3 egg omelettes. In the last hour, 8 customers order the 4 egg omelettes. How many eggs does Theo need to make all the omelettes? \n", + "At Theo’s cafe, he makes 3 egg and 4 egg omelettes. His cafe is open from 7:00 a.m. to 11:00 a.m. In the first hour, 5 customers order the 3 egg omelettes. In the second hour, 7 customers order the 4 egg omelettes. In the third hour, 3 customers order the 3 egg omelettes. In the last hour, 8 customers order the 4 egg omelettes. How many eggs does Theo need to make all the omelettes? \n", "Answer:\n", "84 \n", "Response:\n", @@ -3819,7 +3831,7 @@ "- First week: 40 new followers\n", "- Second week: Half of 40 (which is 20 followers)\n", "- Third week: Half of 20 (which is 10 followers)\n", - "- Total: 100 + 40 + 20 + 1\u6485c\n", + "- Total: 100 + 40 + 20 + 1撅c\n", "\n", "**Sarah's Followers:**\n", "- At the start: 50 followers\n", @@ -3836,7 +3848,7 @@ "- First week: 40 new followers\n", "- Second week: Half of 40 (which is 20 followers)\n", "- Third week: Half of 20 (which is 10 followers)\n", - "- Total: 100 + 40 + 20 + 1\u6485c\n", + "- Total: 100 + 40 + 20 + 1撅c\n", "\n", "**Sarah's Followers:**\n", "- At the start: 50 followers\n", @@ -3850,7 +3862,7 @@ "9 \n", "Response:\n", "\n", - "Jimmy starts with 18 cards. He gives 3 cards to Bob. That leaves Jimmy with 18 - 3 = 15 cards. Then he gives Mary twice as many cards as he gave to Bob. Giving Bob 3 cards, he gives Mary 2 * 3 = 6 cards. Hence, Jimmy\u2019s total number of cards is reduced by 6 more, so he is left with 15 - 6 = 9 cards.\n", + "Jimmy starts with 18 cards. He gives 3 cards to Bob. That leaves Jimmy with 18 - 3 = 15 cards. Then he gives Mary twice as many cards as he gave to Bob. Giving Bob 3 cards, he gives Mary 2 * 3 = 6 cards. Hence, Jimmy’s total number of cards is reduced by 6 more, so he is left with 15 - 6 = 9 cards.\n", "\n", "The final number of cards Jimmy has left is 9.\n", "\n", @@ -3968,7 +3980,7 @@ "\n", "To find the percentage of ads that are not interesting and don't get blocked, we need to break down the problem into several steps:\n", "\n", - "1. First, let\u2019s consider the total percentage of ads that are not blocked by AdBlock. According to the information, AdBlock blocks all but 20% of ads. This means that 80% of the ads get through AdBlock.\n", + "1. First, let’s consider the total percentage of ads that are not blocked by AdBlock. According to the information, AdBlock blocks all but 20% of ads. This means that 80% of the ads get through AdBlock.\n", "\n", "2. Of these 80% ads, 20% of them are interesting. Therefore, 80% of 80% = 80% * (80/100) = 64% are interested and pass through AdBlock.\n", "\n", @@ -3979,7 +3991,7 @@ "\n", "To find the percentage of ads that are not interesting and don't get blocked, we need to break down the problem into several steps:\n", "\n", - "1. First, let\u2019s consider the total percentage of ads that are not blocked by AdBlock. According to the information, AdBlock blocks all but 20% of ads. This means that 80% of the ads get through AdBlock.\n", + "1. First, let’s consider the total percentage of ads that are not blocked by AdBlock. According to the information, AdBlock blocks all but 20% of ads. This means that 80% of the ads get through AdBlock.\n", "\n", "2. Of these 80% ads, 20% of them are interesting. Therefore, 80% of 80% = 80% * (80/100) = 64% are interested and pass through AdBlock.\n", "\n", @@ -4054,7 +4066,7 @@ "\n", "To determine how many points Lizzie, Nathalie, and Aimee scored, we begin by calculating Lizzie's and Nathalie's scores. \n", "- Lizzie scored 4 points. \n", - "- Nathalie scored 3 more points than Lizzie, so Nathalie\u2019s score is:\n", + "- Nathalie scored 3 more points than Lizzie, so Nathalie’s score is:\n", " \\[\n", " 4 + 3 = 7 \\text{ points}\n", " \\]\n", @@ -4064,7 +4076,7 @@ " \\[\n", " 4 + 7 = 11 \\text{ points}\n", " \\]\n", - "- Aimee\u2019s score is twice this combined score:\n", + "- Aimee’s score is twice this combined score:\n", " \\[\n", " 2 \\times 11 = 22 \\text{ points}\n", " \\]\n", @@ -4076,7 +4088,7 @@ "\n", "To determine how many points Lizzie, Nathalie, and Aimee scored, we begin by calculating Lizzie's and Nathalie's scores. \n", "- Lizzie scored 4 points. \n", - "- Nathalie scored 3 more points than Lizzie, so Nathalie\u2019s score is:\n", + "- Nathalie scored 3 more points than Lizzie, so Nathalie’s score is:\n", " \\[\n", " 4 + 3 = 7 \\text{ points}\n", " \\]\n", @@ -4086,7 +4098,7 @@ " \\[\n", " 4 + 7 = 11 \\text{ points}\n", " \\]\n", - "- Aimee\u2019s score is twice this combined score:\n", + "- Aimee’s score is twice this combined score:\n", " \\[\n", " 2 \\times 11 = 22 \\text{ points}\n", " \\]\n", @@ -4356,9 +4368,9 @@ "3. Ken won 5 more games than Dave. Hence, the number of games Ken won is \\(10 + 5 = 15\\).\n", "\n", "Now, we can calculate the total number of games played by summing up the games won by all three:\n", - "- Ken\u2019s games: 15\n", - "- Dave\u2019s games: 10\n", - "- Jerry\u2019s games: 7\n", + "- Ken’s games: 15\n", + "- Dave’s games: 10\n", + "- Jerry’s games: 7\n", "\n", "The total number of games they played is:\n", "\\[ 15 + 10 + 7 = 32 \\]\n", @@ -4373,9 +4385,9 @@ "3. Ken won 5 more games than Dave. Hence, the number of games Ken won is \\(10 + 5 = 15\\).\n", "\n", "Now, we can calculate the total number of games played by summing up the games won by all three:\n", - "- Ken\u2019s games: 15\n", - "- Dave\u2019s games: 10\n", - "- Jerry\u2019s games: 7\n", + "- Ken’s games: 15\n", + "- Dave’s games: 10\n", + "- Jerry’s games: 7\n", "\n", "The total number of games they played is:\n", "\\[ 15 + 10 + 7 = 32 \\]\n", @@ -5132,10 +5144,10 @@ "1 \n", "Response:\n", "\n", - "Duke needs 17 more points to tie the record and ends up breaking the record by 5 points, meaning he scored a total of 257 + 5 = 262 points in the final game. With 5 free throws (1 point each) and 4 regular baskets (2 points each) he scored, this brings his points from these to 5*(1) + 4*(2) = 5 + 8 = 13 points. This means the points he scored beyond the free throws and regular baskets must have been two-pointers or three-pointers, and they totaled 262 - 13 = 249 points. Since normally Duke scores normally 2 three-pointers per game, let\u2019s assume he scored x three-pointers in the final game. Therefore, the total points from three-pointers would be 3*x, and he scored 13 + 249 = \n", + "Duke needs 17 more points to tie the record and ends up breaking the record by 5 points, meaning he scored a total of 257 + 5 = 262 points in the final game. With 5 free throws (1 point each) and 4 regular baskets (2 points each) he scored, this brings his points from these to 5*(1) + 4*(2) = 5 + 8 = 13 points. This means the points he scored beyond the free throws and regular baskets must have been two-pointers or three-pointers, and they totaled 262 - 13 = 249 points. Since normally Duke scores normally 2 three-pointers per game, let’s assume he scored x three-pointers in the final game. Therefore, the total points from three-pointers would be 3*x, and he scored 13 + 249 = \n", "Extracted:\n", "\n", - "Duke needs 17 more points to tie the record and ends up breaking the record by 5 points, meaning he scored a total of 257 + 5 = 262 points in the final game. With 5 free throws (1 point each) and 4 regular baskets (2 points each) he scored, this brings his points from these to 5*(1) + 4*(2) = 5 + 8 = 13 points. This means the points he scored beyond the free throws and regular baskets must have been two-pointers or three-pointers, and they totaled 262 - 13 = 249 points. Since normally Duke scores normally 2 three-pointers per game, let\u2019s assume he scored x three-pointers in the final game. Therefore, the total points from three-pointers would be 3*x, and he scored 13 + 249 =\n", + "Duke needs 17 more points to tie the record and ends up breaking the record by 5 points, meaning he scored a total of 257 + 5 = 262 points in the final game. With 5 free throws (1 point each) and 4 regular baskets (2 points each) he scored, this brings his points from these to 5*(1) + 4*(2) = 5 + 8 = 13 points. This means the points he scored beyond the free throws and regular baskets must have been two-pointers or three-pointers, and they totaled 262 - 13 = 249 points. Since normally Duke scores normally 2 three-pointers per game, let’s assume he scored x three-pointers in the final game. Therefore, the total points from three-pointers would be 3*x, and he scored 13 + 249 =\n", "-------------------- Question:\n", "Corey downloaded two movie series from his Netflix account with 12 and 14 seasons per series, respectively. However, in the week, his computer got a mechanical failure, and he lost two episodes from each season for both series. If each season in the movie series that Corey downloaded had 16 episodes, how many episodes remained after the computer's mechanical failure? \n", "Answer:\n", @@ -5317,7 +5329,7 @@ "- Cost for ten oranges: $46 \\times 10 = $460\n", "- Cost for ten mangoes: $57.50 \\times\n", "-------------------- Question:\n", - "The lights in Malcolm\u2019s house are flickering, and he hopes that replacing all of his white lights with colored lights will make it stop. He buys 12 red lights, 3 times as many blue lights, and 6 green lights. If he still has 5 colored lights left to buy, how many white lights did Malcolm have initially? \n", + "The lights in Malcolm’s house are flickering, and he hopes that replacing all of his white lights with colored lights will make it stop. He buys 12 red lights, 3 times as many blue lights, and 6 green lights. If he still has 5 colored lights left to buy, how many white lights did Malcolm have initially? \n", "Answer:\n", "59 \n", "Response:\n", @@ -6067,30 +6079,30 @@ "Extracted:\n", "$37500\n", "-------------------- Question:\n", - "36 liters of diesel fuel is worth \u20ac18. The tank of this pickup truck can hold 64 liters. How much does a full tank of diesel fuel cost? \n", + "36 liters of diesel fuel is worth €18. The tank of this pickup truck can hold 64 liters. How much does a full tank of diesel fuel cost? \n", "Answer:\n", "32 \n", "Response:\n", "\n", - "To determine the cost of a full tank of diesel fuel, we first need to identify the cost per liter of diesel fuel. Given that 36 liters of fuel cost \u20ac18, we can find the cost per liter by dividing the total cost by the number of liters:\n", + "To determine the cost of a full tank of diesel fuel, we first need to identify the cost per liter of diesel fuel. Given that 36 liters of fuel cost €18, we can find the cost per liter by dividing the total cost by the number of liters:\n", "\n", "\\[\n", - "\\text{Cost per liter} = \\frac{\\text{Total Cost}}{\\text{Total Liters}} = \\frac{18 \\, \\text{\u20ac}}{36 \\, \\text{L}} = 0.5 \\, \\text{\u20ac/L}\n", + "\\text{Cost per liter} = \\frac{\\text{Total Cost}}{\\text{Total Liters}} = \\frac{18 \\, \\text{€}}{36 \\, \\text{L}} = 0.5 \\, \\text{€/L}\n", "\\]\n", "\n", - "Now that we know the cost per liter is \u20ac0.5, we can find out how much 64 liters would cost by multiplying the cost per liter by the total number of liters the truck's tank can hold:\n", + "Now that we know the cost per liter is €0.5, we can find out how much 64 liters would cost by multiplying the cost per liter by the total number of liters the truck's tank can hold:\n", "\n", "\\[\n", "\\text{Cost for 64 liters} = 64 \\, \\text{L} \\times 0.5 \\, \n", "Extracted:\n", "\n", - "To determine the cost of a full tank of diesel fuel, we first need to identify the cost per liter of diesel fuel. Given that 36 liters of fuel cost \u20ac18, we can find the cost per liter by dividing the total cost by the number of liters:\n", + "To determine the cost of a full tank of diesel fuel, we first need to identify the cost per liter of diesel fuel. Given that 36 liters of fuel cost €18, we can find the cost per liter by dividing the total cost by the number of liters:\n", "\n", "\\[\n", - "\\text{Cost per liter} = \\frac{\\text{Total Cost}}{\\text{Total Liters}} = \\frac{18 \\, \\text{\u20ac}}{36 \\, \\text{L}} = 0.5 \\, \\text{\u20ac/L}\n", + "\\text{Cost per liter} = \\frac{\\text{Total Cost}}{\\text{Total Liters}} = \\frac{18 \\, \\text{€}}{36 \\, \\text{L}} = 0.5 \\, \\text{€/L}\n", "\\]\n", "\n", - "Now that we know the cost per liter is \u20ac0.5, we can find out how much 64 liters would cost by multiplying the cost per liter by the total number of liters the truck's tank can hold:\n", + "Now that we know the cost per liter is €0.5, we can find out how much 64 liters would cost by multiplying the cost per liter by the total number of liters the truck's tank can hold:\n", "\n", "\\[\n", "\\text{Cost for 64 liters} = 64 \\, \\text{L} \\times 0.5 \\,\n", @@ -6241,7 +6253,7 @@ "\n", "To determine the total value of Smaug's hoard expressed as a number of copper coins, we need to consider the value of each type of coin in terms of copper coins, and then sum them up.\n", "\n", - "- First, let\u2019s convert the silver coins to copper coins. Given each silver coin is worth 8 copper coins, 60 silver coins will be worth:\n", + "- First, let’s convert the silver coins to copper coins. Given each silver coin is worth 8 copper coins, 60 silver coins will be worth:\n", " \\[\n", " 60 \\text{ silver coins} \\times 8 \\text{ copper coins per silver coin} = 480 \\text{ copper coins}\n", " \\]\n", @@ -6256,7 +6268,7 @@ "\n", "To determine the total value of Smaug's hoard expressed as a number of copper coins, we need to consider the value of each type of coin in terms of copper coins, and then sum them up.\n", "\n", - "- First, let\u2019s convert the silver coins to copper coins. Given each silver coin is worth 8 copper coins, 60 silver coins will be worth:\n", + "- First, let’s convert the silver coins to copper coins. Given each silver coin is worth 8 copper coins, 60 silver coins will be worth:\n", " \\[\n", " 60 \\text{ silver coins} \\times 8 \\text{ copper coins per silver coin} = 480 \\text{ copper coins}\n", " \\]\n", @@ -6378,7 +6390,7 @@ "Extracted:\n", "54\n", - "Tate initially has 32 tickets. He buys 2 more tickets, so his total number of tickets becomes 32 + 2 = 34. We need to determine Peyton\u2019s number of tickets. His friend Peyton has half as many tickets as Tate. So, Peyton has 34 / 2 = 17 tickets. Together, Tate and Peyton have 34 + 17 = 51 tickets.\n", + "Tate initially has 32 tickets. He buys 2 more tickets, so his total number of tickets becomes 32 + 2 = 34. We need to determine Peyton’s number of tickets. His friend Peyton has half as many tickets as Tate. So, Peyton has 34 / 2 = 17 tickets. Together, Tate and Peyton have 34 + 17 = 51 tickets.\n", "\n", "51 \n", "Extracted:\n", @@ -6638,7 +6650,7 @@ "15 \n", "Response:\n", "\n", - "The total weight of swords that need to be transported is 1200 pounds. There are 10 squads and each squad consists of 8 orcs. Therefore, the total number of orcs is 10 * 8 = 80 orcs. To find out how many pounds of swords each orc needs to carry, divide the total weight by the total number of orcs. Thus, 1200 pounds \u00f7 80 orcs = 15 pounds per orc. \n", + "The total weight of swords that need to be transported is 1200 pounds. There are 10 squads and each squad consists of 8 orcs. Therefore, the total number of orcs is 10 * 8 = 80 orcs. To find out how many pounds of swords each orc needs to carry, divide the total weight by the total number of orcs. Thus, 1200 pounds ÷ 80 orcs = 15 pounds per orc. \n", "\n", "15 \n", "Extracted:\n", @@ -6655,7 +6667,7 @@ "\n", "Since each round around the island takes 30 minutes, calculating the total time spent is straightforward. It requires multiplying the total number of rounds by the time it takes to complete one rotation.\n", "\n", - "Total time = 26 rounds \u00d7 30 minutes per round = 780 minutes.\n", + "Total time = 26 rounds × 30 minutes per round = 780 minutes.\n", "\n", "Thus, the Polar Bears spent 780 minutes circling the island over that weekend.\n", "\n", @@ -6964,8 +6976,8 @@ "\n", "Kenzo originally has 80 office chairs with 5 legs each and 20 round tables with 3 legs each. So, the total number of legs from the chairs and tables is calculated as follows: \n", "\n", - "Number of legs from chairs = 80 chairs \u00d7 5 legs per chair = 400 legs\n", - "Number of legs from tables = 20 tables \u00d7 3 legs per table = 60 legs\n", + "Number of legs from chairs = 80 chairs × 5 legs per chair = 400 legs\n", + "Number of legs from tables = 20 tables × 3 legs per table = 60 legs\n", "\n", "So, the total number of legs initially is: 400 legs + 60 legs = 460 legs\n", "\n", @@ -6976,8 +6988,8 @@ "\n", "Kenzo originally has 80 office chairs with 5 legs each and 20 round tables with 3 legs each. So, the total number of legs from the chairs and tables is calculated as follows: \n", "\n", - "Number of legs from chairs = 80 chairs \u00d7 5 legs per chair = 400 legs\n", - "Number of legs from tables = 20 tables \u00d7 3 legs per table = 60 legs\n", + "Number of legs from chairs = 80 chairs × 5 legs per chair = 400 legs\n", + "Number of legs from tables = 20 tables × 3 legs per table = 60 legs\n", "\n", "So, the total number of legs initially is: 400 legs + 60 legs = 460 legs\n", "\n", @@ -7013,7 +7025,7 @@ "\n", "The total number of people the bus carried\n", "-------------------- Question:\n", - "Sharon\u2019s vacation rental has a Keurig coffee machine. She will be there for 40 days. She has 3 cups of coffee (3 coffee pods) every morning. Her coffee pods come 30 pods to a box for $8.00. How much will she spend on coffee to last her for the entire vacation? \n", + "Sharon’s vacation rental has a Keurig coffee machine. She will be there for 40 days. She has 3 cups of coffee (3 coffee pods) every morning. Her coffee pods come 30 pods to a box for $8.00. How much will she spend on coffee to last her for the entire vacation? \n", "Answer:\n", "32 \n", "Response:\n", @@ -7059,7 +7071,7 @@ "12 \n", "Response:\n", "\n", - "To find out how many of the computers sold are desktops, we first calculate the number of laptops and netbooks sold based on the given fractions. Mr. Lu\u2019s store sold a total of 72 computers. \n", + "To find out how many of the computers sold are desktops, we first calculate the number of laptops and netbooks sold based on the given fractions. Mr. Lu’s store sold a total of 72 computers. \n", "\n", "Since half of their sales are laptops, we can calculate the number of laptops sold as:\n", "\\[ \\text{Number of laptops} = \\frac{1}{2} \\times 72 = 36 \\]\n", @@ -7071,7 +7083,7 @@ "\\[ \\text{Number of desktops} = 72 - (\\text{Number of laptops} + \\text{Number of netbooks \n", "Extracted:\n", "\n", - "To find out how many of the computers sold are desktops, we first calculate the number of laptops and netbooks sold based on the given fractions. Mr. Lu\u2019s store sold a total of 72 computers. \n", + "To find out how many of the computers sold are desktops, we first calculate the number of laptops and netbooks sold based on the given fractions. Mr. Lu’s store sold a total of 72 computers. \n", "\n", "Since half of their sales are laptops, we can calculate the number of laptops sold as:\n", "\\[ \\text{Number of laptops} = \\frac{1}{2} \\times 72 = 36 \\]\n", @@ -7250,7 +7262,7 @@ "12 \n", "Response:\n", "\n", - "According to the problem, Ken's house is twice as far from Dawn's house as Mary's house is from Dawn's house. We are given that Ken's house is 4 miles away from Dawn's house. Given this information, we can deduce that Mary's house is 2 miles away from Dawn\u2019s house (since Ken\u2019s house is twice as far from Dawn's house as Mary's house is).\n", + "According to the problem, Ken's house is twice as far from Dawn's house as Mary's house is from Dawn's house. We are given that Ken's house is 4 miles away from Dawn's house. Given this information, we can deduce that Mary's house is 2 miles away from Dawn’s house (since Ken’s house is twice as far from Dawn's house as Mary's house is).\n", "\n", "First, we calculate the distance Ken travels from his house to Dawn's house:\n", "\\[ 4 \\text{ miles} \\]\n", @@ -7258,13 +7270,13 @@ "Next, Ken travels from Dawn's house to Mary's house, which is 2 miles:\n", "\\[ 2 \\text{ miles} \\]\n", "\n", - "Then, Ken travels back from Mary\u2019s house to Dawn\u2019s house, which is 2 miles:\n", + "Then, Ken travels back from Mary’s house to Dawn’s house, which is 2 miles:\n", "\\[ 2 \\text{ miles} \\]\n", "\n", - "Finally, Ken travels from Dawn\u2019s house back to his own house. By the problem statement, Ken's house is 4 miles away from \n", + "Finally, Ken travels from Dawn’s house back to his own house. By the problem statement, Ken's house is 4 miles away from \n", "Extracted:\n", "\n", - "According to the problem, Ken's house is twice as far from Dawn's house as Mary's house is from Dawn's house. We are given that Ken's house is 4 miles away from Dawn's house. Given this information, we can deduce that Mary's house is 2 miles away from Dawn\u2019s house (since Ken\u2019s house is twice as far from Dawn's house as Mary's house is).\n", + "According to the problem, Ken's house is twice as far from Dawn's house as Mary's house is from Dawn's house. We are given that Ken's house is 4 miles away from Dawn's house. Given this information, we can deduce that Mary's house is 2 miles away from Dawn’s house (since Ken’s house is twice as far from Dawn's house as Mary's house is).\n", "\n", "First, we calculate the distance Ken travels from his house to Dawn's house:\n", "\\[ 4 \\text{ miles} \\]\n", @@ -7272,10 +7284,10 @@ "Next, Ken travels from Dawn's house to Mary's house, which is 2 miles:\n", "\\[ 2 \\text{ miles} \\]\n", "\n", - "Then, Ken travels back from Mary\u2019s house to Dawn\u2019s house, which is 2 miles:\n", + "Then, Ken travels back from Mary’s house to Dawn’s house, which is 2 miles:\n", "\\[ 2 \\text{ miles} \\]\n", "\n", - "Finally, Ken travels from Dawn\u2019s house back to his own house. By the problem statement, Ken's house is 4 miles away from\n", + "Finally, Ken travels from Dawn’s house back to his own house. By the problem statement, Ken's house is 4 miles away from\n", "-------------------- Question:\n", "Wilson goes to a fast-food restaurant. He buys 2 hamburgers for $5 each and 3 bottles of cola for $2 each. Wilson uses his $4 discount coupon. How much money does he pay in total? \n", "Answer:\n", @@ -7283,9 +7295,9 @@ "Response:\n", "\n", "Wilson buys 2 hamburgers at $5 each, so the total for hamburgers is:\n", - "2 \u00d7 $5 = $10\n", + "2 × $5 = $10\n", "He buys 3 bottles of cola at $2 each, so the total for cola is:\n", - "3 \u00d7 $2 = $6\n", + "3 × $2 = $6\n", "Adding both the total amount for hamburgers and cola gives us:\n", "$10 + $6 = $16\n", "After using his $4 discount coupon, the total amount he has to pay is:\n", @@ -7469,7 +7481,7 @@ "Extracted:\n", "230\n", "-------------------- Question:\n", - "Michael\u2019s largest watermelon weighs 8 pounds. His neighbor, Clay, grew a watermelon three times that size. And their mutual friend, John, grew a watermelon that was half the size of Clay\u2019s. How big was John\u2019s watermelon? \n", + "Michael’s largest watermelon weighs 8 pounds. His neighbor, Clay, grew a watermelon three times that size. And their mutual friend, John, grew a watermelon that was half the size of Clay’s. How big was John’s watermelon? \n", "Answer:\n", "12 \n", "Response:\n", @@ -7837,7 +7849,7 @@ "1. **Possible combinations with pepperoni and without peppers:**\n", " - Meat: 4 (excluding pepperoni)\n", " - Vegetable: 4 (excluding peppers)\n", - " - Total combinations with this setup = 4 (meat) \u00d7 4 (vegetable) = 16\n", + " - Total combinations with this setup = 4 (meat) × 4 (vegetable) = 16\n", "\n", "2. **Possible combinations without pepperoni:**\n", " \n", @@ -7852,7 +7864,7 @@ "1. **Possible combinations with pepperoni and without peppers:**\n", " - Meat: 4 (excluding pepperoni)\n", " - Vegetable: 4 (excluding peppers)\n", - " - Total combinations with this setup = 4 (meat) \u00d7 4 (vegetable) = 16\n", + " - Total combinations with this setup = 4 (meat) × 4 (vegetable) = 16\n", "\n", "2. **Possible combinations without pepperoni:**\n", "-------------------- Question:\n", @@ -7960,7 +7972,7 @@ "12 \n", "Response:\n", "\n", - "The total number of people at the reunion was the sum of kids and adults, which is 45 kids + 123 adults = 168 people. Since there are 14 tables, to find out how many people sat at each table, we divide the total number of people by the number of tables: 168 people \u00f7 14 tables = 12 people per table.\n", + "The total number of people at the reunion was the sum of kids and adults, which is 45 kids + 123 adults = 168 people. Since there are 14 tables, to find out how many people sat at each table, we divide the total number of people by the number of tables: 168 people ÷ 14 tables = 12 people per table.\n", "\n", "12 \n", "Extracted:\n", @@ -8072,7 +8084,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Processed prompts: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 1/1 [00:00<00:00, 1.71it/s, est. speed input: 63.38 toks/s, output: 25.69 toks/s]\n" + "Processed prompts: 100%|██████████| 1/1 [00:00<00:00, 1.71it/s, est. speed input: 63.38 toks/s, output: 25.69 toks/s]\n" ] }, { @@ -8154,7 +8166,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Processed prompts: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 1/1 [00:03<00:00, 3.06s/it, est. speed input: 14.05 toks/s, output: 29.09 toks/s]\n" + "Processed prompts: 100%|██████████| 1/1 [00:03<00:00, 3.06s/it, est. speed input: 14.05 toks/s, output: 29.09 toks/s]\n" ] }, { @@ -8306,7 +8318,7 @@ " \n", " \n", "\n", - " Join Discord if you need help + \u2b50\ufe0f Star us on Github \u2b50\ufe0f\n", + " Join Discord if you need help + ⭐️ Star us on Github ⭐️\n", "\n" ] } @@ -8673,9 +8685,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_9258ed2d94924de3b944105e280fbf5f", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_1c68fbbc833d4c3fbbb686f15b9ba939", - "value": "special_tokens_map.json:\u2007100%" + "value": "special_tokens_map.json: 100%" } }, "0cc22299405f4361b1b5f6234f35bae3": { @@ -8861,9 +8873,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_d6bd8a9af6e84500956bb565f2fe2c4e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_6643bd32eb984a929159c8d441530aaf", - "value": "\u2007605/605\u2007[00:00<00:00,\u200752.3kB/s]" + "value": " 605/605 [00:00<00:00, 52.3kB/s]" } }, "0f9dcd9b698b4af699cfafdf2874612c": { @@ -8882,9 +8894,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_f8198301b2db406685189589e59efd9a", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e813e26451e3476a82a9b75668b638a1", - "value": "\u20072.36G/2.36G\u2007[00:17<00:00,\u2007511MB/s]" + "value": " 2.36G/2.36G [00:17<00:00, 511MB/s]" } }, "1060550e9dec4fa59e2b9b488af61b48": { @@ -9074,9 +9086,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_6a49aa5a2b3646d4b772e2398eacf72d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_edeb0f283b0b495eab6e06a15fa38af0", - "value": "added_tokens.json:\u2007100%" + "value": "added_tokens.json: 100%" } }, "16e05d45084c44b68697316bc8791b88": { @@ -9119,9 +9131,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_989519f3e61b4132b3f88fbdf29703a9", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e0f87f6fd7c84261818624defe575a1c", - "value": "Map:\u2007100%" + "value": "Map: 100%" } }, "1793da35014840b08d776b4abc9a1212": { @@ -9140,9 +9152,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_231be59acb2a4921ba53ba97bdf5888a", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_5aa0e699560c4fe89250b4ccac6401cb", - "value": "train-00000-of-00001.parquet:\u2007100%" + "value": "train-00000-of-00001.parquet: 100%" } }, "1a619a0a53c149c3b9b0cbd53f1d2034": { @@ -9161,9 +9173,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c626fa7f5c624ec5a414e6f58d71ea96", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_2a8fb4b980ab45be81ae39d5c6b7bad4", - "value": "\u20077.94k/7.94k\u2007[00:00<00:00,\u2007445kB/s]" + "value": " 7.94k/7.94k [00:00<00:00, 445kB/s]" } }, "1b1e49195c5b48a1b6e4364db01c5c87": { @@ -9927,9 +9939,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_784a5c0d201d46459a160e0dc8857fd9", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_d753c1ecc6034237ba4401a007451bfb", - "value": "\u20071.67M/1.67M\u2007[00:00<00:00,\u20071.96MB/s]" + "value": " 1.67M/1.67M [00:00<00:00, 1.96MB/s]" } }, "2f9a64ad318248bf99fccdedafaddfc0": { @@ -9948,9 +9960,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_7c67cf4d83e34cd6a475405be45f345e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_c184e417ad134e94995a55d4bc7c94b5", - "value": "Generating\u2007train\u2007split:\u2007100%" + "value": "Generating train split: 100%" } }, "32719fe47c8543f69bcddc9b9bc0f719": { @@ -10015,9 +10027,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_5cf88ae3c2fc4419a71eb45f28f67384", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_327fa24e70ee4cae903fe5673bcf3bfc", - "value": "vocab.json:\u2007100%" + "value": "vocab.json: 100%" } }, "32defd049d8a473c8ad8fc954c0e3d13": { @@ -10036,9 +10048,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_694494752ab2458ead83b0331af20122", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_1cc89fa4d7c744c69c3be1a3d0b3ba2d", - "value": "tokenizer_config.json:\u2007100%" + "value": "tokenizer_config.json: 100%" } }, "32f42609c389450bbaad10f0a884c527": { @@ -10125,7 +10137,7 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c314f7a0c6084790a2620ec231297525", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_950c1b85a8bd456ab928df4cc3710b21", "value": "" } @@ -10523,9 +10535,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_b245d8e4c115438583ef8b433d4d16be", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_df7f9d244f70440aaf401c6a856a6ed0", - "value": "merges.txt:\u2007100%" + "value": "merges.txt: 100%" } }, "4379f29ff4a74df784bb2fb3989c64cc": { @@ -10544,9 +10556,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_62cbedb2d67d4a2c8888b46e56e39f65", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_57f2c072b0574f74a7dbe2a9d303cbd4", - "value": "\u20071.67M/1.67M\u2007[00:00<00:00,\u20072.32MB/s]" + "value": " 1.67M/1.67M [00:00<00:00, 2.32MB/s]" } }, "4382289897b54e63b22541d0b45c9bf3": { @@ -10605,9 +10617,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c6b2d407256549a5bc7543cb6d7475b4", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_545251749b10408d82c9a60b914dce34", - "value": "special_tokens_map.json:\u2007100%" + "value": "special_tokens_map.json: 100%" } }, "44b6ecc9b9fc436e92701ec2409e0dd1": { @@ -10626,9 +10638,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_96519710bf5047239f6800100471ee6e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e9b020ce2dda442b871b41e417210b52", - "value": "\u20071319/1319\u2007[00:00<00:00,\u200747778.62\u2007examples/s]" + "value": " 1319/1319 [00:00<00:00, 47778.62 examples/s]" } }, "451d181b9ef24b0e9fc3914a4466b9f1": { @@ -11026,9 +11038,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_e14a085ca7e6410ca78fe0f27533be6d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_6d204a683e2d49deb5f504f91ee72044", - "value": "tokenizer_config.json:\u2007100%" + "value": "tokenizer_config.json: 100%" } }, "589a3857411f49d4ae019465f536da45": { @@ -11203,9 +11215,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_11bd7b89fa794a899db9c83cb2a99df5", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_0e9d3d2244a54eb781311c01965c33e4", - "value": "\u20077473/7473\u2007[00:00<00:00,\u20078101.99\u2007examples/s]" + "value": " 7473/7473 [00:00<00:00, 8101.99 examples/s]" } }, "5cda6bd66e3248c4be477f1c1b40198d": { @@ -12007,9 +12019,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_78774135f2d1468d9172fa7588afef20", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_9b3233c52ecc4b259deed2d5d6f332eb", - "value": "Loading\u2007safetensors\u2007checkpoint\u2007shards:\u2007100%\u2007Completed\u2007|\u20071/1\u2007[00:01<00:00,\u2007\u20071.41s/it]\n" + "value": "Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:01<00:00,  1.41s/it]\n" } }, "7529f4d84b9f46709ec5f0e22b75baf4": { @@ -12080,9 +12092,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_269b131121554544a21d47596a24aa58", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_73758a3ca9c14d7eaa6504d8b9b6e333", - "value": "added_tokens.json:\u2007100%" + "value": "added_tokens.json: 100%" } }, "78345adf6d8143e3a649a05c5d476115": { @@ -12604,9 +12616,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_b131891a909743f5826429a84a74b4f9", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_3745da58146d445abe863a248e86d9be", - "value": "README.md:\u2007100%" + "value": "README.md: 100%" } }, "8a89212807f243d4b135a3188f4fb6d0": { @@ -12625,9 +12637,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c9b781db7ee2422b8dc4eb6aa2575d87", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_5100eb13df1c453980942ec78d0e810b", - "value": "\u200711.4M/11.4M\u2007[00:00<00:00,\u200740.1MB/s]" + "value": " 11.4M/11.4M [00:00<00:00, 40.1MB/s]" } }, "8aa00f7a2de64af8b0a5b8cd5fefb14e": { @@ -12684,9 +12696,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_52ba857bdfc742a2aae9e9a949054e5c", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_589a3857411f49d4ae019465f536da45", - "value": "\u20077.36k/7.36k\u2007[00:00<00:00,\u2007512kB/s]" + "value": " 7.36k/7.36k [00:00<00:00, 512kB/s]" } }, "8ee15a3d02954d31899d0b6a8dbf7988": { @@ -12802,9 +12814,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1ccb12282661417383a7f96bd5ea4eac", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_32a16b0cbbe548f4869cf665d737b9ba", - "value": "model.safetensors:\u2007100%" + "value": "model.safetensors: 100%" } }, "950c1b85a8bd456ab928df4cc3710b21": { @@ -13010,9 +13022,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_3b6d4b41992b43acac5daa31f119cd30", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_c5b4960fc3ba4459a6eaff43772050a3", - "value": "\u2007605/605\u2007[00:00<00:00,\u200748.4kB/s]" + "value": " 605/605 [00:00<00:00, 48.4kB/s]" } }, "989519f3e61b4132b3f88fbdf29703a9": { @@ -13083,9 +13095,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a422e26291bb47d08e36b7d94fffcc9b", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_799dc8fe6f7d4e4db68ebd6b1d8fdbd5", - "value": "vocab.json:\u2007100%" + "value": "vocab.json: 100%" } }, "9929aaffe7614506ae934a14d2b5c3ca": { @@ -13171,9 +13183,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_29a3f867c8bd4769b8e649b8912776a2", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_2317c9a94dfe4a62a8605038ebc2210a", - "value": "\u2007271/271\u2007[00:00<00:00,\u200715.4kB/s]" + "value": " 271/271 [00:00<00:00, 15.4kB/s]" } }, "9bcb0000da6e434d9b241a09d9b46244": { @@ -13294,9 +13306,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_420154f0bbe344a7979769a8ef066df7", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_1ee34b9171604805887aee3baa4ff099", - "value": "\u20072.31M/2.31M\u2007[00:00<00:00,\u200725.1MB/s]" + "value": " 2.31M/2.31M [00:00<00:00, 25.1MB/s]" } }, "a422e26291bb47d08e36b7d94fffcc9b": { @@ -13443,9 +13455,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_5d190801b10b49a698c3a69c7cd1f253", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_0112cfe0a27e43fd9509ef8da050e188", - "value": "\u2007614/614\u2007[00:00<00:00,\u200752.1kB/s]" + "value": " 614/614 [00:00<00:00, 52.1kB/s]" } }, "a8501b8551d4431fbb708d6556e7dd85": { @@ -13583,7 +13595,7 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_24757877452d48f0afe92bb66e2c48ac", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_b64be6a0d6df44899ba6fe0a47f676e1", "value": "" } @@ -13830,9 +13842,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_37e2f38471d4481eb93031ca52bdb785", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_067aaf24d87a400b9f636762051598f3", - "value": "Loading\u2007safetensors\u2007checkpoint\u2007shards:\u2007100%\u2007Completed\u2007|\u20071/1\u2007[00:02<00:00,\u2007\u20072.58s/it]\n" + "value": "Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:02<00:00,  2.58s/it]\n" } }, "b64be6a0d6df44899ba6fe0a47f676e1": { @@ -13866,9 +13878,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4ddc8fa0c9834cd3b36274a9e0e05504", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_ec13219478284bbdbe69cbcd3dff5c34", - "value": "\u2007614/614\u2007[00:00<00:00,\u200741.6kB/s]" + "value": " 614/614 [00:00<00:00, 41.6kB/s]" } }, "bc81dfa869f94f059545a7d59b7eb70a": { @@ -13924,9 +13936,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_71898e427b1245d682178a90c6752e8e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_90295f1308054c66b557da231d25d3d5", - "value": "\u2007419k/419k\u2007[00:00<00:00,\u200721.7MB/s]" + "value": " 419k/419k [00:00<00:00, 21.7MB/s]" } }, "c13ff304d7bc4c67a1f51da5d2303f96": { @@ -14378,9 +14390,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c404db8298364742b8ce076cddcb7641", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_03aca6d92597461481419eac9bdb6635", - "value": "tokenizer.json:\u2007100%" + "value": "tokenizer.json: 100%" } }, "c9b781db7ee2422b8dc4eb6aa2575d87": { @@ -14503,9 +14515,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_95fa6a54e7e543c7b2dc5d5cf04daf58", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_8ee15a3d02954d31899d0b6a8dbf7988", - "value": "Generating\u2007test\u2007split:\u2007100%" + "value": "Generating test split: 100%" } }, "d13cc9053e7a4982b6fc7f00474f25c7": { @@ -14782,9 +14794,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_d36027362b41413fa3ee4670dd820a88", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_d94e3315b99747c39190fee9b79f00ac", - "value": "\u20077473/7473\u2007[00:00<00:00,\u200716342.30\u2007examples/s]" + "value": " 7473/7473 [00:00<00:00, 16342.30 examples/s]" } }, "df7f9d244f70440aaf401c6a856a6ed0": { @@ -14947,9 +14959,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_aaf79572906c4443b5542bef5dee8055", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_bc81dfa869f94f059545a7d59b7eb70a", - "value": "generation_config.json:\u2007100%" + "value": "generation_config.json: 100%" } }, "e505a51d928542b883743a6d48c5cbd3": { @@ -15246,9 +15258,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1f88a409f3c64ea291861837e830ce39", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e9881edb2c7846fea2070c8e724b25cb", - "value": "\u200711.4M/11.4M\u2007[00:00<00:00,\u200742.5MB/s]" + "value": " 11.4M/11.4M [00:00<00:00, 42.5MB/s]" } }, "edeb0f283b0b495eab6e06a15fa38af0": { @@ -15306,9 +15318,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_0c2e39b20d5941969101b793b94fdbfa", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_f0c74de390814c2b9f38f7cc02458426", - "value": "test-00000-of-00001.parquet:\u2007100%" + "value": "test-00000-of-00001.parquet: 100%" } }, "f0c74de390814c2b9f38f7cc02458426": { @@ -15470,9 +15482,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_9929aaffe7614506ae934a14d2b5c3ca", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_1060550e9dec4fa59e2b9b488af61b48", - "value": "\u20072.78M/2.78M\u2007[00:00<00:00,\u20073.27MB/s]" + "value": " 2.78M/2.78M [00:00<00:00, 3.27MB/s]" } }, "f77e9e785cc04c00822fc8b7ef2959da": { @@ -15611,9 +15623,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_32f42609c389450bbaad10f0a884c527", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_a9e874a076234974a102b02551ec8666", - "value": "\u20077.36k/7.36k\u2007[00:00<00:00,\u2007580kB/s]" + "value": " 7.36k/7.36k [00:00<00:00, 580kB/s]" } }, "f89bfff6526c43bcb75c93f987bfba72": { @@ -15632,9 +15644,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c69730dfeec14cedaf8917ee9ad27dcb", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_00177cac24374c7d9144e3856594ddaf", - "value": "tokenizer.json:\u2007100%" + "value": "tokenizer.json: 100%" } }, "fa996ce2f8f24a5d80914c4ef5b6e788": { @@ -15705,9 +15717,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_878702a019e34789b5b2f68eb321f2c8", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_91998be216844c079144c18278f660d7", - "value": "\u20072.78M/2.78M\u2007[00:01<00:00,\u20072.68MB/s]" + "value": " 2.78M/2.78M [00:01<00:00, 2.68MB/s]" } }, "fc87fbccba304554be90701b39429918": { @@ -15778,9 +15790,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_0e8c4086734c4296884db9c5b67ddd33", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_c3a97f3130bc42c1be2f8c7dc81ca6cc", - "value": "merges.txt:\u2007100%" + "value": "merges.txt: 100%" } }, "fdcbb976f35e4aa383d9ebcd7304849f": { @@ -15857,4 +15869,4 @@ }, "nbformat": 4, "nbformat_minor": 0 -} \ No newline at end of file +} diff --git a/nb/Llama3.1_(8B)-GRPO.ipynb b/nb/Llama3.1_(8B)-GRPO.ipynb index e32de5e..60dc1f8 100644 --- a/nb/Llama3.1_(8B)-GRPO.ipynb +++ b/nb/Llama3.1_(8B)-GRPO.ipynb @@ -8,7 +8,7 @@ "
\n", "\n", "\n", - " Join Discord if you need help + \u2b50 Star us on Github \u2b50\n", + " Join Discord if you need help + ⭐ Star us on Github ⭐\n", "
\n", "\n", "To install Unsloth on your own computer, follow the installation instructions on our Github page [here](https://docs.unsloth.ai/get-started/installing-+-updating).\n", @@ -48,14 +48,40 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\nimport os\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n # If you're not in Colab, just use pip install or uv pip install\n !pip install unsloth vllm\nelse:\n pass # For Colab / Kaggle, we need extra instructions hidden below \\/" + "source": [ + "%%capture\n", + "import os\n", + "if \"COLAB_\" not in \"\".join(os.environ.keys()):\n", + " # If you're not in Colab, just use pip install or uv pip install\n", + " !pip install unsloth vllm\n", + "else:\n", + " pass # For Colab / Kaggle, we need extra instructions hidden below \\/" + ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], - "source": "#@title Colab Extra Install { display-mode: \"form\" }\n%%capture\nimport os\n!pip install --upgrade -qqq uv\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n # If you're not in Colab, just use pip install!\n !pip install unsloth vllm\nelse:\n try: import numpy; get_numpy = f\"numpy=={numpy.__version__}\"\n except: get_numpy = \"numpy\"\n try: import subprocess; is_t4 = \"Tesla T4\" in str(subprocess.check_output([\"nvidia-smi\"]))\n except: is_t4 = False\n get_vllm, get_triton = (\"vllm==0.10.1\", \"triton==3.2.0\") if is_t4 else (\"vllm\", \"triton\")\n !uv pip install -qqq --upgrade \\\n unsloth {get_vllm} {get_numpy} torchvision bitsandbytes xformers\n !uv pip install -qqq {get_triton}\n!uv pip install transformers==4.55.4" + "source": [ + "#@title Colab Extra Install { display-mode: \"form\" }\n", + "%%capture\n", + "import os\n", + "!pip install --upgrade -qqq uv\n", + "if \"COLAB_\" not in \"\".join(os.environ.keys()):\n", + " # If you're not in Colab, just use pip install!\n", + " !pip install unsloth vllm\n", + "else:\n", + " try: import numpy; get_numpy = f\"numpy=={numpy.__version__}\"\n", + " except: get_numpy = \"numpy\"\n", + " try: import subprocess; is_t4 = \"Tesla T4\" in str(subprocess.check_output([\"nvidia-smi\"]))\n", + " except: is_t4 = False\n", + " get_vllm, get_triton = (\"vllm==0.10.1\", \"triton==3.2.0\") if is_t4 else (\"vllm\", \"triton\")\n", + " !uv pip install -qqq --upgrade \\\n", + " unsloth {get_vllm} {get_numpy} torchvision bitsandbytes xformers\n", + " !uv pip install -qqq {get_triton}\n", + "!uv pip install transformers==4.55.4" + ] }, { "cell_type": "markdown", @@ -200,8 +226,8 @@ "output_type": "stream", "text": [ "Unsloth: Patching Xformers to fix some performance issues.\n", - "\ud83e\udda5 Unsloth: Will patch your computer to enable 2x faster free finetuning.\n", - "\ud83e\udda5 Unsloth Zoo will now patch everything to make training faster!\n", + "🦥 Unsloth: Will patch your computer to enable 2x faster free finetuning.\n", + "🦥 Unsloth Zoo will now patch everything to make training faster!\n", "INFO 03-07 09:37:36 __init__.py:207] Automatically detected platform cuda.\n", "==((====))== Unsloth 2025.3.8: Fast Llama patching. Transformers: 4.48.3. vLLM: 0.7.3.\n", " \\\\ /| Tesla T4. Num GPUs = 1. Max memory: 14.741 GB. Platform: Linux.\n", @@ -353,7 +379,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Capturing CUDA graph shapes: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 23/23 [00:43<00:00, 1.88s/it]" + "Capturing CUDA graph shapes: 100%|██████████| 23/23 [00:43<00:00, 1.88s/it]" ] }, { @@ -463,7 +489,7 @@ }, { "cell_type": "code", - "execution_count": 3, + "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", @@ -691,14 +717,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", @@ -4305,7 +4331,7 @@ "\n", "However, we also need to consider the pants Dani already has. Dani initially had 50 pants.\n", "-------------------- Question:\n", - "Jessica\u2019s class is going to a farm for a field trip. The school will bring all 35 students in the class plus 4 adult chaperones. The farm entrance fee for students costs $5 and $6 for adults. How much will the school pay for the farm entrance in all? \n", + "Jessica’s class is going to a farm for a field trip. The school will bring all 35 students in the class plus 4 adult chaperones. The farm entrance fee for students costs $5 and $6 for adults. How much will the school pay for the farm entrance in all? \n", "Answer:\n", "199 \n", "Response:\n", @@ -4690,7 +4716,7 @@ "\n", "So Emmanuel will get 80 jelly beans.\n", "-------------------- Question:\n", - "Libby has 160 quarters in her piggy bank. She has to pay $35 to replace her sister\u2019s dress that she borrowed and ruined. After replacing the dress, how many quarters will Libby have left? \n", + "Libby has 160 quarters in her piggy bank. She has to pay $35 to replace her sister’s dress that she borrowed and ruined. After replacing the dress, how many quarters will Libby have left? \n", "Answer:\n", "20 \n", "Response:\n", @@ -4706,7 +4732,7 @@ "\n", "Since $5 is equal to 5 / 0.25 = 20 quarters. \n", "\n", - "20 quarters will be left after she pays for the sister\u2019s dress.\n", + "20 quarters will be left after she pays for the sister’s dress.\n", "\n", "20 \n", "Extracted:\n", @@ -4933,7 +4959,7 @@ "Extracted:\n", "21\n", "-------------------- Question:\n", - "Steven is preparing a shipment of boxes to deliver to a customer for his chemical supply business. The products are very delicate and must be carefully packed, so partially filled boxes can\u2019t be shipped. Steven has three trucks that can be sent on each delivery. Each truck can carry a load of no more than 2,000 pounds of cargo. Some of the boxes weigh 10 pounds after being packed, and some of the boxes weigh 40 pounds when packed. Steven\u2019s customer has ordered equal quantities of both the lighter and heavier products. How many boxes of products can Steven ship to his customer in each delivery? \n", + "Steven is preparing a shipment of boxes to deliver to a customer for his chemical supply business. The products are very delicate and must be carefully packed, so partially filled boxes can’t be shipped. Steven has three trucks that can be sent on each delivery. Each truck can carry a load of no more than 2,000 pounds of cargo. Some of the boxes weigh 10 pounds after being packed, and some of the boxes weigh 40 pounds when packed. Steven’s customer has ordered equal quantities of both the lighter and heavier products. How many boxes of products can Steven ship to his customer in each delivery? \n", "Answer:\n", "240 \n", "Response:\n", @@ -4945,12 +4971,12 @@ "\n", "Since each truck can carry a maximum of 2,000 pounds, we can set up the following equation:\n", "\n", - "50x \u2264 2000\n", + "50x ≤ 2000\n", "\n", "Now, we can divide both sides by 50:\n", "\n", - "x \u2264 2000 / 50\n", - "x \u2264 40\n", + "x ≤ 2000 / 50\n", + "x ≤ 40\n", "\n", "Therefore, Steven can ship at most 40 boxes of products per delivery to his customer.\n", "\n", @@ -5168,7 +5194,7 @@ "\n", "Since there are 6 people splitting the cost evenly, the cost per person can be calculated by dividing the total cost by 6:\n", "\n", - "$2340 \u00f7 6 = $390\n", + "$2340 ÷ 6 = $390\n", "$390 \n", "Extracted:\n", "$390\n", @@ -5235,7 +5261,7 @@ "Extracted:\n", "39\n", "-------------------- Question:\n", - "There are 25 different layers of rock in a canyon\u2019s strata. Five of the layers are limestone. Half of the rest are sandstone. Four of the remaining are quartz. Half of the rest are shale. The rest are all different kinds of rock. How many kinds of rock make up only one layer in the canyon\u2019s strata? \n", + "There are 25 different layers of rock in a canyon’s strata. Five of the layers are limestone. Half of the rest are sandstone. Four of the remaining are quartz. Half of the rest are shale. The rest are all different kinds of rock. How many kinds of rock make up only one layer in the canyon’s strata? \n", "Answer:\n", "3 \n", "Response:\n", @@ -5537,7 +5563,7 @@ "\n", "Since the actual value from spending the quarters is $400 and the value from melting is $300, he would get $\\frac{300}{400}$ or $\\frac{3}{4}$, or 0.75 times as much money by melting down the quarters than by spending them.\n", "-------------------- Question:\n", - "Ever since she was a little girl, Sheila has dreamed of traveling the world. To help fund her dream, she bought a large piggy bank in December and started saving. By last week, she had saved $3,000. Pleased with her progress, she has decided to continue saving $276 per month, for 4 years. Today, Sheila\u2019s family secretly added $7,000 into the piggy bank. At the end of 4 years, how much money will be in Sheila\u2019s piggy bank? \n", + "Ever since she was a little girl, Sheila has dreamed of traveling the world. To help fund her dream, she bought a large piggy bank in December and started saving. By last week, she had saved $3,000. Pleased with her progress, she has decided to continue saving $276 per month, for 4 years. Today, Sheila’s family secretly added $7,000 into the piggy bank. At the end of 4 years, how much money will be in Sheila’s piggy bank? \n", "Answer:\n", "23248 \n", "Response:\n", @@ -5555,7 +5581,7 @@ "Extracted:\n", "$23,248.\n", "-------------------- Question:\n", - "Tom\u2019s cat is 8 years old. His rabbit is half the age of his cat. His dog is three times as old as his rabbit. How old is the dog? \n", + "Tom’s cat is 8 years old. His rabbit is half the age of his cat. His dog is three times as old as his rabbit. How old is the dog? \n", "Answer:\n", "12 \n", "Response:\n", @@ -5646,7 +5672,7 @@ "Extracted:\n", "$355\n", "-------------------- Question:\n", - "Haman\u2019s father has an egg business supplying the local market. On a Wednesday morning, his father sends him to go and collect 10 trays of eggs for sale from their store. While loading the eggs into the car, he accidentally drops two trays. He calls his father telling him this, and is told to add 7 more trays for sale. How many eggs were sold that day? \n", + "Haman’s father has an egg business supplying the local market. On a Wednesday morning, his father sends him to go and collect 10 trays of eggs for sale from their store. While loading the eggs into the car, he accidentally drops two trays. He calls his father telling him this, and is told to add 7 more trays for sale. How many eggs were sold that day? \n", "Answer:\n", "540 \n", "Response:\n", @@ -5727,7 +5753,7 @@ "Extracted:\n", "28\n", "-------------------- Question:\n", - "All 20 of Grant\u2019s baby teeth have fallen out and he has a total of $54 from the tooth fairy. Every time he lost a tooth, he put it under his pillow for the tooth fairy, except for one that he dropped on the way home from school and another he swallowed accidentally. The tooth fairy left Grant $20 when he lost his first tooth. How much did the tooth fairy leave him per tooth after his first tooth, assuming equal money exchanged for each tooth thereafter? \n", + "All 20 of Grant’s baby teeth have fallen out and he has a total of $54 from the tooth fairy. Every time he lost a tooth, he put it under his pillow for the tooth fairy, except for one that he dropped on the way home from school and another he swallowed accidentally. The tooth fairy left Grant $20 when he lost his first tooth. How much did the tooth fairy leave him per tooth after his first tooth, assuming equal money exchanged for each tooth thereafter? \n", "Answer:\n", "2 \n", "Response:\n", @@ -6043,7 +6069,7 @@ "Extracted:\n", "3\n", "-------------------- Question:\n", - "After Bella\u2019s ballet performance, she received 2 dozen roses from her parents, and 2 roses from each of her 10 dancer friends. How many roses did Bella receive? \n", + "After Bella’s ballet performance, she received 2 dozen roses from her parents, and 2 roses from each of her 10 dancer friends. How many roses did Bella receive? \n", "Answer:\n", "44 \n", "Response:\n", @@ -6235,14 +6261,14 @@ "Response:\n", "To find the average price per movie, we need to calculate the total cost of the movies and then divide it by the total number of movies.\n", "\n", - "The total cost of the DVDs is 8 \u00d7 $12 = $96.\n", - "The total cost of the Blu-rays is 4 \u00d7 $18 = $72.\n", + "The total cost of the DVDs is 8 × $12 = $96.\n", + "The total cost of the Blu-rays is 4 × $18 = $72.\n", "\n", "The total number of movies Chris bought is 8 + 4 = 12.\n", "\n", "The total cost of all the movies is $96 + $72 = $168.\n", "\n", - "The average price per movie is $168 \u00f7 12 = $14.\n", + "The average price per movie is $168 ÷ 12 = $14.\n", "\n", "14 \n", "Extracted:\n", @@ -6465,8 +6491,8 @@ "Response:\n", "To find the price Janice needs to pay for one juice and one sandwich, we need to first find out the price of one juice and the price of one sandwich.\n", "\n", - "One juice costs $10 \u00f7 5 = $2.\n", - "One sandwich costs $6 \u00f7 2 = $3.\n", + "One juice costs $10 ÷ 5 = $2.\n", + "One sandwich costs $6 ÷ 2 = $3.\n", "\n", "So, in total she would need to pay $2 (for the juice) + $3 (for the sandwich) = $5.\n", "\n", @@ -6509,12 +6535,12 @@ "\n", "He bought 9 cans of tuna for $14.50. To find the cost of one can, we need to divide the total amount he paid by the number of cans.\n", "\n", - "$14.50 \u00f7 9 cans = $1.6111 (rounded to four decimal places).\n", + "$14.50 ÷ 9 cans = $1.6111 (rounded to four decimal places).\n", "\n", "To convert this to cents, we multiply by 100:\n", - "$1.6111 \u00d7 100 = 161.11 cents.\n", + "$1.6111 × 100 = 161.11 cents.\n", "\n", - "Since he used 5 coupons for 25 cents each, he saved a total of 5 \u00d7 25 = 125 cents.\n", + "Since he used 5 coupons for 25 cents each, he saved a total of 5 × 25 = 125 cents.\n", "\n", "Therefore, the actual cost of one can without the coupons would be 161.11 - 1.25 (125 cents) = 161.11 - 1.25 = 160 cents.\n", "\n", @@ -6793,7 +6819,7 @@ "Extracted:\n", "220\n", "-------------------- Question:\n", - "Heath spent his weekend helping at his uncle\u2019s farm planting carrots. He planted 400 rows of carrots with 300 plants in each row. He used a machine to get the planting done and it took him 20 hours. How many carrots did he plant each hour? \n", + "Heath spent his weekend helping at his uncle’s farm planting carrots. He planted 400 rows of carrots with 300 plants in each row. He used a machine to get the planting done and it took him 20 hours. How many carrots did he plant each hour? \n", "Answer:\n", "6,000 \n", "Response:\n", @@ -7047,11 +7073,11 @@ "\n", "Now we multiply both sides of the equation by 3/2.\n", "\n", - "I = 20 \u00d7 3/2 \n", + "I = 20 × 3/2 \n", "I = 30\n", "\n", "Since Dina has twice the number of dolls as Ivy, Dina has 2I dolls.\n", - "Dina has 2 \u00d7 30 = 60 dolls.\n", + "Dina has 2 × 30 = 60 dolls.\n", "\n", "60 \n", "Extracted:\n", @@ -7546,7 +7572,7 @@ "\n", "Since Leonard's age must be an integer, we must have a non-integer age for Jerome and/or Nina. We'll check if their ages can be non-integers.\n", "-------------------- Question:\n", - "Adam's father deposited $2000 in the bank. It receives 8% interest paid throughout the year, and he withdraws the interest as soon as it is deposited. How much will Adam\u2019s father have, including his deposit and the interest received after 2 and a half years? \n", + "Adam's father deposited $2000 in the bank. It receives 8% interest paid throughout the year, and he withdraws the interest as soon as it is deposited. How much will Adam’s father have, including his deposit and the interest received after 2 and a half years? \n", "Answer:\n", "2400 \n", "Response:\n", @@ -7749,7 +7775,7 @@ "Extracted:\n", "25\n", "-------------------- Question:\n", - "Louise is in a toy store. She already has 28 toys worth $10 each in her cart. On her way to the till she adds 20 teddy bears to the cart. If the $580 in Louise\u2019s wallet is exactly enough to pay for all the toys, how much does each teddy bear cost? \n", + "Louise is in a toy store. She already has 28 toys worth $10 each in her cart. On her way to the till she adds 20 teddy bears to the cart. If the $580 in Louise’s wallet is exactly enough to pay for all the toys, how much does each teddy bear cost? \n", "Answer:\n", "15 \n", "Response:\n", @@ -8450,7 +8476,7 @@ "Extracted:\n", "1600\n", "-------------------- Question:\n", - "Julian has 80 Facebook friends. 60% are boys and 40% are girls. Boyd has twice as many friends who are girls and has 100 friends total. What percentage of Boyd\u2019s friends are boys? \n", + "Julian has 80 Facebook friends. 60% are boys and 40% are girls. Boyd has twice as many friends who are girls and has 100 friends total. What percentage of Boyd’s friends are boys? \n", "Answer:\n", "36 \n", "Response:\n", @@ -8504,7 +8530,7 @@ "Extracted:\n", "21600\n", "-------------------- Question:\n", - "Claudia offers art classes to kids and charges $10.00 for her one-hour class. If 20 kids attend Saturday\u2019s class and half that many attend Sunday\u2019s class, how much money does she make? \n", + "Claudia offers art classes to kids and charges $10.00 for her one-hour class. If 20 kids attend Saturday’s class and half that many attend Sunday’s class, how much money does she make? \n", "Answer:\n", "300 \n", "Response:\n", @@ -8736,7 +8762,7 @@ "Extracted:\n", "9\n", "-------------------- Question:\n", - "It\u2019s exam season and Tristan has several exams to prepare for. On Monday, he studies for 4 hours then studies for twice this long on Tuesday. On Wednesday, Thursday, and Friday he studies for 3 hours each day. He wants to study for a total of 25 hours over the week and divides the remaining amount of study time evenly between Saturday and Sunday. How many hours does Tristan spend studying on Saturday? \n", + "It’s exam season and Tristan has several exams to prepare for. On Monday, he studies for 4 hours then studies for twice this long on Tuesday. On Wednesday, Thursday, and Friday he studies for 3 hours each day. He wants to study for a total of 25 hours over the week and divides the remaining amount of study time evenly between Saturday and Sunday. How many hours does Tristan spend studying on Saturday? \n", "Answer:\n", "2 \n", "Response:\n", @@ -8900,7 +8926,7 @@ "Extracted:\n", "48\n", "-------------------- Question:\n", - "Daisy\u2019s milk cow provides 16 cups of milk per day. 75% of the milk is consumed by Daisy\u2019s kids. Daisy uses 50% of the remaining milk to cook with. How much milk is left over? \n", + "Daisy’s milk cow provides 16 cups of milk per day. 75% of the milk is consumed by Daisy’s kids. Daisy uses 50% of the remaining milk to cook with. How much milk is left over? \n", "Answer:\n", "2 \n", "Response:\n", @@ -9053,7 +9079,7 @@ "Extracted:\n", "192\n", "-------------------- Question:\n", - "Erika and her 3 siblings are drawing with chalk outside. Another 3 friends join them and ask if they can help with the drawing. Erika loses 2 pieces of chalk as they are counting and the group realizes there isn't enough chalk for everyone. Erika\u2019s mom brings out another 12 pieces of chalk for them to share and there is now enough chalk for everyone to have 3 pieces each. How many pieces of chalk did Erika and her siblings originally have? \n", + "Erika and her 3 siblings are drawing with chalk outside. Another 3 friends join them and ask if they can help with the drawing. Erika loses 2 pieces of chalk as they are counting and the group realizes there isn't enough chalk for everyone. Erika’s mom brings out another 12 pieces of chalk for them to share and there is now enough chalk for everyone to have 3 pieces each. How many pieces of chalk did Erika and her siblings originally have? \n", "Answer:\n", "11 \n", "Response:\n", @@ -9093,7 +9119,7 @@ "\n", "So, Erika and her siblings originally had 11 pieces of chalk.\n", "-------------------- Question:\n", - "Brian is a real estate agent. He makes a 2% commission on each sale. During June, he sold three houses. The houses sold for $157,000, $499,000, and $125,000. What was Brian\u2019s total commission on these three sales? \n", + "Brian is a real estate agent. He makes a 2% commission on each sale. During June, he sold three houses. The houses sold for $157,000, $499,000, and $125,000. What was Brian’s total commission on these three sales? \n", "Answer:\n", "15,620 \n", "Response:\n", @@ -9200,7 +9226,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Processed prompts: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 1/1 [00:51<00:00, 51.81s/it, est. speed input: 0.75 toks/s, output: 17.01 toks/s]\n" + "Processed prompts: 100%|██████████| 1/1 [00:51<00:00, 51.81s/it, est. speed input: 0.75 toks/s, output: 17.01 toks/s]\n" ] }, { @@ -9209,7 +9235,7 @@ "type": "string" }, "text/plain": [ - "'**Calculating Pi using Python**\\n\\nPi (\u03c0) is a mathematical constant representing the ratio of a circle\\'s circumference to its diameter. Here\\'s a simple and efficient way to calculate an approximation of pi using Python.\\n\\n### Using the Monte Carlo Method\\n\\nThe Monte Carlo method is a computational algorithm that uses random sampling to approximate a value. In this case, we can use it to estimate pi by generating random points within a square and checking if they fall inside a quarter-circle inscribed within it.\\n\\n```python\\nimport random\\nimport math\\n\\ndef estimate_pi(num_samples):\\n \"\"\"\\n Estimate the value of pi using the Monte Carlo method.\\n\\n Args:\\n num_samples (int): The number of random points to generate.\\n\\n Returns:\\n float: An approximation of pi.\\n \"\"\"\\n points_inside_circle = 0\\n\\n for _ in range(num_samples):\\n x, y = random.random(), random.random()\\n distance = x**2 + y**2\\n if distance <= 1:\\n points_inside_circle += 1\\n\\n return (points_inside_circle / num_samples) * 4\\n\\n# Example usage:\\nnum_samples = 1000000\\napprox_pi = estimate_pi(num_samples)\\nprint(f\"Approximation of pi with {num_samples} samples: {approx_pi}\")\\n```\\n\\nThis code generates `num_samples` random points within the square (-1, -1) to (1, 1) and checks if each point falls inside the quarter-circle. The ratio of points inside the circle to the total number of samples multiplied by 4 gives an approximation of pi.\\n\\n### Using the Bailey-Borwein-Plouffe Formula\\n\\nThe Bailey-Borwein-Plouffe (BBP) formula is another way to calculate pi. It\\'s a spigot algorithm that uses a series expansion to generate the digits of pi.\\n\\n```python\\ndef bbp_pi(n):\\n \"\"\"\\n Calculate the Bailey-Borwein-Plouffe formula for pi.\\n\\n Args:\\n n (int): The number of terms to use in the series expansion.\\n\\n Returns:\\n float: An approximation of pi.\\n \"\"\"\\n pi = 0.0\\n for k in range(n):\\n pi += (1 / (16 ** k)) * (\\n 4 / (8 * k + 1) -\\n 2 / (8 * k + 4) -\\n 1 / (8 * k + 5) -\\n 1 / (8 * k + 6)\\n )\\n return pi\\n\\n# Example usage:\\nn = 100\\napprox_pi = bbp_pi(n)\\nprint(f\"Approximation of pi with {n} terms: {approx_pi}\")\\n```\\n\\nThis code calculates an approximation of pi using the BBP formula with `n` terms.\\n\\n### Using the Gauss-Legendre Algorithm\\n\\nThe Gauss-Legendre algorithm is another method for calculating pi. It\\'s a recursive algorithm that uses an iterative process to improve the estimate of pi.\\n\\n```python\\ndef gauss_legendre(n):\\n \"\"\"\\n Calculate pi using the Gauss-Legendre algorithm.\\n\\n Args:\\n n (int): The number of iterations.\\n\\n Returns:\\n float: An approximation of pi.\\n \"\"\"\\n a, b = 1.0, (1 + math.sqrt(5)) / 2.0\\n t = (1 + math.sqrt(5)) / 4.0\\n p = 1.0\\n for _ in range(n):\\n a, b = (a + b) / 2.0, math.sqrt(a * b)\\n t -= p * (a - b)\\n p *= 2.0\\n a = a + t / p\\n b = math.sqrt(b - t / p)\\n return a * b * 16.0\\n\\n# Example usage:\\nn = 100\\napprox_pi = gauss_legendre(n)\\nprint(f\"Approximation of pi with {n} iterations: {approx_pi}\")\\n```\\n\\nThis code calculates an approximation of pi using the Gauss-Legendre algorithm with `n` iterations.\\n\\nRemember that these are simplified examples and may not be the most efficient or accurate methods for calculating pi. For more accurate calculations, you can use specialized libraries or software.'" + "'**Calculating Pi using Python**\\n\\nPi (π) is a mathematical constant representing the ratio of a circle\\'s circumference to its diameter. Here\\'s a simple and efficient way to calculate an approximation of pi using Python.\\n\\n### Using the Monte Carlo Method\\n\\nThe Monte Carlo method is a computational algorithm that uses random sampling to approximate a value. In this case, we can use it to estimate pi by generating random points within a square and checking if they fall inside a quarter-circle inscribed within it.\\n\\n```python\\nimport random\\nimport math\\n\\ndef estimate_pi(num_samples):\\n \"\"\"\\n Estimate the value of pi using the Monte Carlo method.\\n\\n Args:\\n num_samples (int): The number of random points to generate.\\n\\n Returns:\\n float: An approximation of pi.\\n \"\"\"\\n points_inside_circle = 0\\n\\n for _ in range(num_samples):\\n x, y = random.random(), random.random()\\n distance = x**2 + y**2\\n if distance <= 1:\\n points_inside_circle += 1\\n\\n return (points_inside_circle / num_samples) * 4\\n\\n# Example usage:\\nnum_samples = 1000000\\napprox_pi = estimate_pi(num_samples)\\nprint(f\"Approximation of pi with {num_samples} samples: {approx_pi}\")\\n```\\n\\nThis code generates `num_samples` random points within the square (-1, -1) to (1, 1) and checks if each point falls inside the quarter-circle. The ratio of points inside the circle to the total number of samples multiplied by 4 gives an approximation of pi.\\n\\n### Using the Bailey-Borwein-Plouffe Formula\\n\\nThe Bailey-Borwein-Plouffe (BBP) formula is another way to calculate pi. It\\'s a spigot algorithm that uses a series expansion to generate the digits of pi.\\n\\n```python\\ndef bbp_pi(n):\\n \"\"\"\\n Calculate the Bailey-Borwein-Plouffe formula for pi.\\n\\n Args:\\n n (int): The number of terms to use in the series expansion.\\n\\n Returns:\\n float: An approximation of pi.\\n \"\"\"\\n pi = 0.0\\n for k in range(n):\\n pi += (1 / (16 ** k)) * (\\n 4 / (8 * k + 1) -\\n 2 / (8 * k + 4) -\\n 1 / (8 * k + 5) -\\n 1 / (8 * k + 6)\\n )\\n return pi\\n\\n# Example usage:\\nn = 100\\napprox_pi = bbp_pi(n)\\nprint(f\"Approximation of pi with {n} terms: {approx_pi}\")\\n```\\n\\nThis code calculates an approximation of pi using the BBP formula with `n` terms.\\n\\n### Using the Gauss-Legendre Algorithm\\n\\nThe Gauss-Legendre algorithm is another method for calculating pi. It\\'s a recursive algorithm that uses an iterative process to improve the estimate of pi.\\n\\n```python\\ndef gauss_legendre(n):\\n \"\"\"\\n Calculate pi using the Gauss-Legendre algorithm.\\n\\n Args:\\n n (int): The number of iterations.\\n\\n Returns:\\n float: An approximation of pi.\\n \"\"\"\\n a, b = 1.0, (1 + math.sqrt(5)) / 2.0\\n t = (1 + math.sqrt(5)) / 4.0\\n p = 1.0\\n for _ in range(n):\\n a, b = (a + b) / 2.0, math.sqrt(a * b)\\n t -= p * (a - b)\\n p *= 2.0\\n a = a + t / p\\n b = math.sqrt(b - t / p)\\n return a * b * 16.0\\n\\n# Example usage:\\nn = 100\\napprox_pi = gauss_legendre(n)\\nprint(f\"Approximation of pi with {n} iterations: {approx_pi}\")\\n```\\n\\nThis code calculates an approximation of pi using the Gauss-Legendre algorithm with `n` iterations.\\n\\nRemember that these are simplified examples and may not be the most efficient or accurate methods for calculating pi. For more accurate calculations, you can use specialized libraries or software.'" ] }, "execution_count": 6, @@ -9282,7 +9308,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Processed prompts: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 1/1 [00:23<00:00, 23.17s/it, est. speed input: 2.63 toks/s, output: 15.80 toks/s]\n" + "Processed prompts: 100%|██████████| 1/1 [00:23<00:00, 23.17s/it, est. speed input: 2.63 toks/s, output: 15.80 toks/s]\n" ] }, { @@ -9434,7 +9460,7 @@ " \n", " \n", "\n", - " Join Discord if you need help + \u2b50\ufe0f Star us on Github \u2b50\ufe0f\n", + " Join Discord if you need help + ⭐️ Star us on Github ⭐️\n", "\n" ] } @@ -9487,9 +9513,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fda3ed1250d34b12aae8401287322548", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e85f549d21dd4c3293f8239a053b0ad4", - "value": "test-00000-of-00001.parquet:\u2007100%" + "value": "test-00000-of-00001.parquet: 100%" } }, "0722b569fa2e45f9af7d4fb8134e632b": { @@ -9560,9 +9586,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_ef2a701c9c594d4d9ffc8379fa9b5899", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_22c31da7c0564f09bbdb04efdefad11a", - "value": "\u20077473/7473\u2007[00:00<00:00,\u20079130.98\u2007examples/s]" + "value": " 7473/7473 [00:00<00:00, 9130.98 examples/s]" } }, "09fca7e3b90f4b17a2344b67100e26bf": { @@ -9603,9 +9629,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fa155f662d1c40f496daef290ff7a5ff", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_9b2a88f9712d409bbf74bf0c2e14df16", - "value": "\u200717.2M/17.2M\u2007[00:00<00:00,\u2007165MB/s]" + "value": " 17.2M/17.2M [00:00<00:00, 165MB/s]" } }, "0b9dd1c421f14c56986cf30abe0a9455": { @@ -9624,9 +9650,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_73177ee1d64c4f56aa393c45c42e7499", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_98aabe0e4167414d943ff3d7759093ad", - "value": "\u200755.5k/55.5k\u2007[00:00<00:00,\u20073.84MB/s]" + "value": " 55.5k/55.5k [00:00<00:00, 3.84MB/s]" } }, "0bc45e643a9b4a0e8db9f85180c108b2": { @@ -9954,9 +9980,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_d4eeb17c3a124c919058c0ffe72a60b6", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_8b6bd37f64454464b44691ae3bef0b9f", - "value": "Loading\u2007safetensors\u2007checkpoint\u2007shards:\u2007100%\u2007Completed\u2007|\u20071/1\u2007[00:27<00:00,\u200727.33s/it]\n" + "value": "Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:27<00:00, 27.33s/it]\n" } }, "18d2fd154ffc4671b76f00760eb35571": { @@ -9990,9 +10016,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_594a583be494419c845fd873904d4d6d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_ba21e0c745714031aae12b1665847e54", - "value": "tokenizer.json:\u2007100%" + "value": "tokenizer.json: 100%" } }, "1a090b75a50b459a8d2f822efc692dfa": { @@ -10191,9 +10217,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1a090b75a50b459a8d2f822efc692dfa", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_0da6c9fa4eeb4b4d832f8bbc614bc8d9", - "value": "\u20077.94k/7.94k\u2007[00:00<00:00,\u2007557kB/s]" + "value": " 7.94k/7.94k [00:00<00:00, 557kB/s]" } }, "22c31da7c0564f09bbdb04efdefad11a": { @@ -10346,9 +10372,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_364d9414fd5f4280be4ad621874e15cb", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_7b3ad76804bd47a68fba0bf9f967cb04", - "value": "\u20075.96G/5.96G\u2007[00:45<00:00,\u2007545MB/s]" + "value": " 5.96G/5.96G [00:45<00:00, 545MB/s]" } }, "2aade13ff2ca4250ae02b7a12b1ef16d": { @@ -10882,9 +10908,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_79da5ea5195c4c63b0ec1ee354511a87", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_a5f3704359df48e4b4c5b418a5f72021", - "value": "\u2007454/454\u2007[00:00<00:00,\u200724.6kB/s]" + "value": " 454/454 [00:00<00:00, 24.6kB/s]" } }, "397a49fc365b4c3a95db709b71fc32e3": { @@ -10955,9 +10981,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_0ea62e14a2d44b8ca7c6f15bca8e1a2b", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_27e78ef30ff64be5af79b6431816f569", - "value": "\u2007239/239\u2007[00:00<00:00,\u200723.8kB/s]" + "value": " 239/239 [00:00<00:00, 23.8kB/s]" } }, "436282df7bcc45c4b951de60f2e4ec53": { @@ -11028,9 +11054,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_269f762b5af041eab750226febe70e07", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_f38c687b78224c8aa2227914998b4627", - "value": "train-00000-of-00001.parquet:\u2007100%" + "value": "train-00000-of-00001.parquet: 100%" } }, "48403d21996b4ebfa931cf260af6c283": { @@ -11153,9 +11179,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fe9dbcf4e85646cb8c684d82252b93e2", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_d36f0dda703c47cfae6f544e7e655806", - "value": "\u200755.5k/55.5k\u2007[00:00<00:00,\u20074.95MB/s]" + "value": " 55.5k/55.5k [00:00<00:00, 4.95MB/s]" } }, "4c69d53cf33c4865af8feead04b810ad": { @@ -11444,9 +11470,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_5e9d2ad532a8447a904688629a91f171", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_2df38622a5ab46979ee53efeb8c31b91", - "value": "README.md:\u2007100%" + "value": "README.md: 100%" } }, "5de480828775465bb76b674ecd713bd0": { @@ -11569,9 +11595,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a67ca6e6c91244978b6b468257ffefd7", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_781f0652012b424aa43cd39028e3a99a", - "value": "generation_config.json:\u2007100%" + "value": "generation_config.json: 100%" } }, "638c59bf00ba4536ad157975faa26984": { @@ -11668,9 +11694,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_81f43f522faa4dcf9e557793a2027e90", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_ae46c8f17c0644fdbd1aa6d412c322b7", - "value": "Generating\u2007train\u2007split:\u2007100%" + "value": "Generating train split: 100%" } }, "6be20c9739de44ffb7ef2465ad17424d": { @@ -11728,9 +11754,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_bc92bf99bcf345b59699eb8f0423f7cd", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_f589e115f63d4fdb95816dccb34c3e3e", - "value": "tokenizer_config.json:\u2007100%" + "value": "tokenizer_config.json: 100%" } }, "70780d08fb00445da32b28b40d33412f": { @@ -12055,9 +12081,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_ff032a23470845ba95a12227baf0e0ce", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_18d2fd154ffc4671b76f00760eb35571", - "value": "\u200717.2M/17.2M\u2007[00:00<00:00,\u2007103MB/s]" + "value": " 17.2M/17.2M [00:00<00:00, 103MB/s]" } }, "7b3ad76804bd47a68fba0bf9f967cb04": { @@ -12091,7 +12117,7 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_7083c21ed3dd4ed6b442afdf4a18a05d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_6e4ce08c86f24beeb63d38d436143dc9", "value": "" } @@ -12188,9 +12214,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_cbe9c4e2f8bc49dfb1d8a4cddc36d724", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_f33aef89be7641a694fb11e3101c7d85", - "value": "\u2007454/454\u2007[00:00<00:00,\u200739.2kB/s]" + "value": " 454/454 [00:00<00:00, 39.2kB/s]" } }, "813e8e40df6447c18e47414ab38f69a2": { @@ -12209,9 +12235,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_115a8f3c523c4cd3aaa472af87fe0e4a", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_26cde13f97db4ae4a0900fb5184a51a2", - "value": "special_tokens_map.json:\u2007100%" + "value": "special_tokens_map.json: 100%" } }, "81f43f522faa4dcf9e557793a2027e90": { @@ -12334,9 +12360,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_32fe42f8a11d4131a3a529b554f028fe", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_39c33d513fb74712be8fcf73093a0c9e", - "value": "\u20071319/1319\u2007[00:00<00:00,\u200728401.73\u2007examples/s]" + "value": " 1319/1319 [00:00<00:00, 28401.73 examples/s]" } }, "8b6bd37f64454464b44691ae3bef0b9f": { @@ -12553,9 +12579,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_d89eae07da0e4344909900e24e3d0d09", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_b1e00fc06b664fc59d4921388f187269", - "value": "Map:\u2007100%" + "value": "Map: 100%" } }, "93f708b722194750a9f66a2597987033": { @@ -12812,9 +12838,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_b095d3eac38847c0b087ffcc5876e0b6", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_9ebd8c869bc144cf8142375e7d7f04dd", - "value": "tokenizer.json:\u2007100%" + "value": "tokenizer.json: 100%" } }, "9ec592e894b945cea7c2a756a96eb959": { @@ -12886,9 +12912,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fac3506146244766b886c715ea57ec7b", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_d068e96c067e43ffa1777a666b0df147", - "value": "tokenizer_config.json:\u2007100%" + "value": "tokenizer_config.json: 100%" } }, "a5f3704359df48e4b4c5b418a5f72021": { @@ -13078,9 +13104,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_436282df7bcc45c4b951de60f2e4ec53", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_911d088df0894243b69fd4f69c46e7f3", - "value": "\u2007419k/419k\u2007[00:00<00:00,\u20078.72MB/s]" + "value": " 419k/419k [00:00<00:00, 8.72MB/s]" } }, "ae46c8f17c0644fdbd1aa6d412c322b7": { @@ -13392,9 +13418,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4c69d53cf33c4865af8feead04b810ad", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_9dbc67661c614d8a82d0790d9b694e6d", - "value": "Loading\u2007safetensors\u2007checkpoint\u2007shards:\u2007100%\u2007Completed\u2007|\u20071/1\u2007[00:14<00:00,\u200714.38s/it]\n" + "value": "Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:14<00:00, 14.38s/it]\n" } }, "c3ec8a75007846519d40f6726d2a01ee": { @@ -13820,9 +13846,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_2aade13ff2ca4250ae02b7a12b1ef16d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_23531535f2fc4b99a5ecd1b63fb93d80", - "value": "\u20072.31M/2.31M\u2007[00:00<00:00,\u200746.9MB/s]" + "value": " 2.31M/2.31M [00:00<00:00, 46.9MB/s]" } }, "d4eeb17c3a124c919058c0ffe72a60b6": { @@ -13945,9 +13971,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_2df204a0982b466eb3c25bf80645b6c1", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_9ec592e894b945cea7c2a756a96eb959", - "value": "special_tokens_map.json:\u2007100%" + "value": "special_tokens_map.json: 100%" } }, "d7e3f2c2441743089ee7ed10eb0319dd": { @@ -13966,9 +13992,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_2c83de81fc91497db859d1be6860d6f4", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_397a49fc365b4c3a95db709b71fc32e3", - "value": "model.safetensors:\u2007100%" + "value": "model.safetensors: 100%" } }, "d89eae07da0e4344909900e24e3d0d09": { @@ -14123,9 +14149,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_cc6097b25ec74decb51a397ccd9fcf27", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_d2af3c19e6ec46a3915f6badf7062b7b", - "value": "\u20077473/7473\u2007[00:00<00:00,\u20072457.10\u2007examples/s]" + "value": " 7473/7473 [00:00<00:00, 2457.10 examples/s]" } }, "e1e8c601a2bb4c0b8f0cdf3bf0dd0b2e": { @@ -14250,7 +14276,7 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fe4ecfe3522c432b964608f137cfb936", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e1e8c601a2bb4c0b8f0cdf3bf0dd0b2e", "value": "" } @@ -14558,9 +14584,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_76dc9524ee764741a1615de6bc43492e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_78201e7953b647ee9656fc7e8f4abbcb", - "value": "Generating\u2007test\u2007split:\u2007100%" + "value": "Generating test split: 100%" } }, "fa155f662d1c40f496daef290ff7a5ff": { @@ -14933,4 +14959,4 @@ }, "nbformat": 4, "nbformat_minor": 0 -} \ No newline at end of file +} diff --git a/nb/Mistral_v0.3_(7B)-GRPO.ipynb b/nb/Mistral_v0.3_(7B)-GRPO.ipynb index c665bc7..d0bf29d 100644 --- a/nb/Mistral_v0.3_(7B)-GRPO.ipynb +++ b/nb/Mistral_v0.3_(7B)-GRPO.ipynb @@ -543,14 +543,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", diff --git a/nb/Phi_4_(14B)-GRPO.ipynb b/nb/Phi_4_(14B)-GRPO.ipynb index c301b75..d17c86d 100644 --- a/nb/Phi_4_(14B)-GRPO.ipynb +++ b/nb/Phi_4_(14B)-GRPO.ipynb @@ -8,7 +8,7 @@ "
\n", "\n", "\n", - " Join Discord if you need help + \u2b50 Star us on Github \u2b50\n", + " Join Discord if you need help + ⭐ Star us on Github ⭐\n", "
\n", "\n", "To install Unsloth on your own computer, follow the installation instructions on our Github page [here](https://docs.unsloth.ai/get-started/installing-+-updating).\n", @@ -48,14 +48,40 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\nimport os\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n # If you're not in Colab, just use pip install or uv pip install\n !pip install unsloth vllm\nelse:\n pass # For Colab / Kaggle, we need extra instructions hidden below \\/" + "source": [ + "%%capture\n", + "import os\n", + "if \"COLAB_\" not in \"\".join(os.environ.keys()):\n", + " # If you're not in Colab, just use pip install or uv pip install\n", + " !pip install unsloth vllm\n", + "else:\n", + " pass # For Colab / Kaggle, we need extra instructions hidden below \\/" + ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], - "source": "#@title Colab Extra Install { display-mode: \"form\" }\n%%capture\nimport os\n!pip install --upgrade -qqq uv\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n # If you're not in Colab, just use pip install!\n !pip install unsloth vllm\nelse:\n try: import numpy; get_numpy = f\"numpy=={numpy.__version__}\"\n except: get_numpy = \"numpy\"\n try: import subprocess; is_t4 = \"Tesla T4\" in str(subprocess.check_output([\"nvidia-smi\"]))\n except: is_t4 = False\n get_vllm, get_triton = (\"vllm==0.10.1\", \"triton==3.2.0\") if is_t4 else (\"vllm\", \"triton\")\n !uv pip install -qqq --upgrade \\\n unsloth {get_vllm} {get_numpy} torchvision bitsandbytes xformers\n !uv pip install -qqq {get_triton}\n!uv pip install transformers==4.55.4" + "source": [ + "#@title Colab Extra Install { display-mode: \"form\" }\n", + "%%capture\n", + "import os\n", + "!pip install --upgrade -qqq uv\n", + "if \"COLAB_\" not in \"\".join(os.environ.keys()):\n", + " # If you're not in Colab, just use pip install!\n", + " !pip install unsloth vllm\n", + "else:\n", + " try: import numpy; get_numpy = f\"numpy=={numpy.__version__}\"\n", + " except: get_numpy = \"numpy\"\n", + " try: import subprocess; is_t4 = \"Tesla T4\" in str(subprocess.check_output([\"nvidia-smi\"]))\n", + " except: is_t4 = False\n", + " get_vllm, get_triton = (\"vllm==0.10.1\", \"triton==3.2.0\") if is_t4 else (\"vllm\", \"triton\")\n", + " !uv pip install -qqq --upgrade \\\n", + " unsloth {get_vllm} {get_numpy} torchvision bitsandbytes xformers\n", + " !uv pip install -qqq {get_triton}\n", + "!uv pip install transformers==4.55.4" + ] }, { "cell_type": "markdown", @@ -182,7 +208,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Capturing CUDA graph shapes: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 19/19 [00:50<00:00, 2.67s/it]" + "Capturing CUDA graph shapes: 100%|██████████| 19/19 [00:50<00:00, 2.67s/it]" ] }, { @@ -315,14 +341,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", @@ -1498,11 +1524,11 @@ "\n", "1. **Calculate the cost of the books:**\n", " - Lynne bought 7 books about cats and 2 books about the solar system. This is a total of 7 + 2 = 9 books.\n", - " - Each book costs $7. So, the total cost for books is 9 books \u00d7 $7/book = $63.\n", + " - Each book costs $7. So, the total cost for books is 9 books × $7/book = $63.\n", "\n", "2. **Calculate the cost of the magazines:**\n", " - Lynne bought 3 magazines.\n", - " - Each magazine costs $4. So, the total cost for magazines is 3 magazines \u00d7 $4/magazine = $12.\n", + " - Each magazine costs $4. So, the total cost for magazines is 3 magazines × $4/magazine = $12.\n", "\n", "3. **Calculate the total amount spent:**\n", " - Add the total cost of the books and the magazines: $63 (books) + $12 (magazines) = $75.\n", @@ -1514,11 +1540,11 @@ "\n", "1. **Calculate the cost of the books:**\n", " - Lynne bought 7 books about cats and 2 books about the solar system. This is a total of 7 + 2 = 9 books.\n", - " - Each book costs $7. So, the total cost for books is 9 books \u00d7 $7/book = $63.\n", + " - Each book costs $7. So, the total cost for books is 9 books × $7/book = $63.\n", "\n", "2. **Calculate the cost of the magazines:**\n", " - Lynne bought 3 magazines.\n", - " - Each magazine costs $4. So, the total cost for magazines is 3 magazines \u00d7 $4/magazine = $12.\n", + " - Each magazine costs $4. So, the total cost for magazines is 3 magazines × $4/magazine = $12.\n", "\n", "3. **Calculate the total amount spent:**\n", " - Add the total cost of the books and the magazines: $63 (books) + $12 (magazines) = $75.\n", @@ -1656,7 +1682,7 @@ "Extracted:\n", "21\n", "-------------------- Question:\n", - "There are 84 people waiting in line to ride a roller coaster at an amusement park.\u00a0 The roller coaster has 7 cars, and each car seats 2 people.\u00a0 How many times will the ride operator have to run the roller coaster to give everyone in line a turn? \n", + "There are 84 people waiting in line to ride a roller coaster at an amusement park.  The roller coaster has 7 cars, and each car seats 2 people.  How many times will the ride operator have to run the roller coaster to give everyone in line a turn? \n", "Answer:\n", "6 \n", "Response:\n", @@ -1738,17 +1764,17 @@ "1. **Necklaces:**\n", " - Price per necklace: $25.00\n", " - Number sold: 5\n", - " - Total from necklaces = 5 necklaces \u00d7 $25.00/necklace = $125.00\n", + " - Total from necklaces = 5 necklaces × $25.00/necklace = $125.00\n", "\n", "2. **Bracelets:**\n", " - Price per bracelet: $15.00\n", " - Number sold: 10\n", - " - Total from bracelets = 10 bracelets \u00d7 $15.00/bracelet = $150.00\n", + " - Total from bracelets = 10 bracelets × $15.00/bracelet = $150.00\n", "\n", "3. **Earrings:**\n", " - Price per pair of earrings: $10.00\n", " - Number sold: 20\n", - " - Total from earrings = 20 pairs \u00d7 $10.00/pair = $200.00\n", + " - Total from earrings = 20 pairs × $10.00/pair = $200.00\n", "\n", "4. **Complete Jewelry Ensembles:**\n", " - Price per ensemble: \n", @@ -1759,17 +1785,17 @@ "1. **Necklaces:**\n", " - Price per necklace: $25.00\n", " - Number sold: 5\n", - " - Total from necklaces = 5 necklaces \u00d7 $25.00/necklace = $125.00\n", + " - Total from necklaces = 5 necklaces × $25.00/necklace = $125.00\n", "\n", "2. **Bracelets:**\n", " - Price per bracelet: $15.00\n", " - Number sold: 10\n", - " - Total from bracelets = 10 bracelets \u00d7 $15.00/bracelet = $150.00\n", + " - Total from bracelets = 10 bracelets × $15.00/bracelet = $150.00\n", "\n", "3. **Earrings:**\n", " - Price per pair of earrings: $10.00\n", " - Number sold: 20\n", - " - Total from earrings = 20 pairs \u00d7 $10.00/pair = $200.00\n", + " - Total from earrings = 20 pairs × $10.00/pair = $200.00\n", "\n", "4. **Complete Jewelry Ensembles:**\n", " - Price per ensemble:\n", @@ -1838,7 +1864,7 @@ " Convert 0.5 to a fraction with a common denominator: \n", " \\( 0.5 = \\frac{3}{6} = \\frac{4}{6} = \\frac{2}{3}\n", "-------------------- Question:\n", - "At Hank\u2019s cafe, he sells big stack pancakes which have 5 pancakes and short stack pancakes which have 3 pancakes. If 9 customers order the short stack and 6 customers order the big stack, how many pancakes does Hank need to make? \n", + "At Hank’s cafe, he sells big stack pancakes which have 5 pancakes and short stack pancakes which have 3 pancakes. If 9 customers order the short stack and 6 customers order the big stack, how many pancakes does Hank need to make? \n", "Answer:\n", "57 \n", "Response:\n", @@ -2101,8 +2127,8 @@ "To determine how many candies are in each packet, we first need to find out how many candies Bobby eats in 3 weeks.\n", "\n", "Bobby's candy consumption breakdown for one week is as follows:\n", - "- Monday to Friday: 2 candies per day \u00d7 5 days = 10 candies\n", - "- Saturday and Sunday: 1 candy per day \u00d7 2 days = 2 candies\n", + "- Monday to Friday: 2 candies per day × 5 days = 10 candies\n", + "- Saturday and Sunday: 1 candy per day × 2 days = 2 candies\n", "\n", "Therefore, in one week, Bobby eats:\n", "\\[ 10 \\text{ candies} + 2 \\text{ candies} = 12 \\text{ candies} \\]\n", @@ -2116,8 +2142,8 @@ "To determine how many candies are in each packet, we first need to find out how many candies Bobby eats in 3 weeks.\n", "\n", "Bobby's candy consumption breakdown for one week is as follows:\n", - "- Monday to Friday: 2 candies per day \u00d7 5 days = 10 candies\n", - "- Saturday and Sunday: 1 candy per day \u00d7 2 days = 2 candies\n", + "- Monday to Friday: 2 candies per day × 5 days = 10 candies\n", + "- Saturday and Sunday: 1 candy per day × 2 days = 2 candies\n", "\n", "Therefore, in one week, Bobby eats:\n", "\\[ 10 \\text{ candies} + 2 \\text{ candies} = 12 \\text{ candies} \\]\n", @@ -2709,7 +2735,7 @@ "Extracted:\n", "20 counselors\n", "-------------------- Question:\n", - "After Bella\u2019s ballet performance, she received 2 dozen roses from her parents, and 2 roses from each of her 10 dancer friends. How many roses did Bella receive? \n", + "After Bella’s ballet performance, she received 2 dozen roses from her parents, and 2 roses from each of her 10 dancer friends. How many roses did Bella receive? \n", "Answer:\n", "44 \n", "Response:\n", @@ -3737,16 +3763,16 @@ " \\[\n", " 2M - 1 = 15 \\quad \\text{(Multiplying both sides by 3)}\n", "-------------------- Question:\n", - "Clyde and Grace are building block towers. Grace\u2019s tower is 8 times the size of Clyde\u2019s at 40 inches tall. How many inches taller is Grace\u2019s tower than Clyde\u2019s? \n", + "Clyde and Grace are building block towers. Grace’s tower is 8 times the size of Clyde’s at 40 inches tall. How many inches taller is Grace’s tower than Clyde’s? \n", "Answer:\n", "35 \n", "Response:\n", "\n", "To solve this problem, we need to determine the height of Clyde's tower and compare it to the height of Grace's tower.\n", "\n", - "1. We know that Grace\u2019s tower is 8 times the size of Clyde\u2019s tower. Let \\( x \\) be the height of Clyde's tower. Therefore, Grace's tower is \\( 8x \\).\n", + "1. We know that Grace’s tower is 8 times the size of Clyde’s tower. Let \\( x \\) be the height of Clyde's tower. Therefore, Grace's tower is \\( 8x \\).\n", "\n", - "2. We are given that Grace\u2019s tower is 40 inches tall. So, we can write the equation:\n", + "2. We are given that Grace’s tower is 40 inches tall. So, we can write the equation:\n", " \\[\n", " 8x = 40\n", " \\]\n", @@ -3758,14 +3784,14 @@ "\n", "4. Clyde's tower is 5 inches tall.\n", "\n", - "5. To find how many inches taller Grace\u2019s tower is compared to Clyde's, we subtract the height of Clyde\u2019s tower \n", + "5. To find how many inches taller Grace’s tower is compared to Clyde's, we subtract the height of Clyde’s tower \n", "Extracted:\n", "\n", "To solve this problem, we need to determine the height of Clyde's tower and compare it to the height of Grace's tower.\n", "\n", - "1. We know that Grace\u2019s tower is 8 times the size of Clyde\u2019s tower. Let \\( x \\) be the height of Clyde's tower. Therefore, Grace's tower is \\( 8x \\).\n", + "1. We know that Grace’s tower is 8 times the size of Clyde’s tower. Let \\( x \\) be the height of Clyde's tower. Therefore, Grace's tower is \\( 8x \\).\n", "\n", - "2. We are given that Grace\u2019s tower is 40 inches tall. So, we can write the equation:\n", + "2. We are given that Grace’s tower is 40 inches tall. So, we can write the equation:\n", " \\[\n", " 8x = 40\n", " \\]\n", @@ -3777,7 +3803,7 @@ "\n", "4. Clyde's tower is 5 inches tall.\n", "\n", - "5. To find how many inches taller Grace\u2019s tower is compared to Clyde's, we subtract the height of Clyde\u2019s tower\n", + "5. To find how many inches taller Grace’s tower is compared to Clyde's, we subtract the height of Clyde’s tower\n", "-------------------- Question:\n", "Andrew's father buys a package of 100 masks. Andrew lives with his 2 parents and 2 siblings. All members of Andrew's family change masks every 4 days. How many days will it take to finish the pack of masks? \n", "Answer:\n", @@ -3793,7 +3819,7 @@ "2. Each member of the family changes their mask every 4 days. This implies that each person uses 1 mask every 4 days.\n", "\n", "3. Therefore, in a 4-day period, all 5 family members together would use:\n", - " - 5 people \u00d7 1 mask per person = 5 masks.\n", + " - 5 people × 1 mask per person = 5 masks.\n", "\n", "4. Now, we want to know how many total days it will take for the package of 100 masks to be used up. Let's calculate how many 4-day periods it takes to use 100 masks:\n", " - Number of 4-day \n", @@ -3808,7 +3834,7 @@ "2. Each member of the family changes their mask every 4 days. This implies that each person uses 1 mask every 4 days.\n", "\n", "3. Therefore, in a 4-day period, all 5 family members together would use:\n", - " - 5 people \u00d7 1 mask per person = 5 masks.\n", + " - 5 people × 1 mask per person = 5 masks.\n", "\n", "4. Now, we want to know how many total days it will take for the package of 100 masks to be used up. Let's calculate how many 4-day periods it takes to use 100 masks:\n", " - Number of 4-day\n", @@ -4047,7 +4073,7 @@ "\n", "Given that \\( x = 210 \\), and we know that after cutting these \\( x \\) bananas\n", "-------------------- Question:\n", - "At Theo\u2019s cafe, he makes 3 egg and 4 egg omelettes. His cafe is open from 7:00 a.m. to 11:00 a.m. In the first hour, 5 customers order the 3 egg omelettes. In the second hour, 7 customers order the 4 egg omelettes. In the third hour, 3 customers order the 3 egg omelettes. In the last hour, 8 customers order the 4 egg omelettes. How many eggs does Theo need to make all the omelettes? \n", + "At Theo’s cafe, he makes 3 egg and 4 egg omelettes. His cafe is open from 7:00 a.m. to 11:00 a.m. In the first hour, 5 customers order the 3 egg omelettes. In the second hour, 7 customers order the 4 egg omelettes. In the third hour, 3 customers order the 3 egg omelettes. In the last hour, 8 customers order the 4 egg omelettes. How many eggs does Theo need to make all the omelettes? \n", "Answer:\n", "84 \n", "Response:\n", @@ -4436,7 +4462,7 @@ "\\]\n", "\n", "Now, calculate the number of slices the waiter receives:\n", - "- Waiter's slices = 8 parts \u00d7 6 slices per part = 48 slices\n", + "- Waiter's slices = 8 parts × 6 slices per part = 48 slices\n", "\n", "Finally, we need to find \"twenty less than the number of slices the waiter ate\":\n", "- Slices the waiter ate = 48\n", @@ -4456,7 +4482,7 @@ "\\]\n", "\n", "Now, calculate the number of slices the waiter receives:\n", - "- Waiter's slices = 8 parts \u00d7 6 slices per part = 48 slices\n", + "- Waiter's slices = 8 parts × 6 slices per part = 48 slices\n", "\n", "Finally, we need to find \"twenty less than the number of slices the waiter ate\":\n", "- Slices the waiter ate = 48\n", @@ -4973,7 +4999,7 @@ "2. **Calculate the weekly equivalent of the monthly housing benefit reduction:**\n", " - The monthly reduction in his housing benefit is $60.\n", " - To find the equivalent weekly reduction, divide the monthly reduction by the average number of weeks in a month (approximately 4.33 weeks per month).\n", - " - Weekly reduction in housing benefit = $60/month \u00f7 4.33 weeks/month \u2248 $13.86/week.\n", + " - Weekly reduction in housing benefit = $60/month ÷ 4.33 weeks/month ≈ $13.86/week.\n", "\n", "3. **Calculate the net weekly effect:**\n", " - Net weekly \n", @@ -4989,7 +5015,7 @@ "2. **Calculate the weekly equivalent of the monthly housing benefit reduction:**\n", " - The monthly reduction in his housing benefit is $60.\n", " - To find the equivalent weekly reduction, divide the monthly reduction by the average number of weeks in a month (approximately 4.33 weeks per month).\n", - " - Weekly reduction in housing benefit = $60/month \u00f7 4.33 weeks/month \u2248 $13.86/week.\n", + " - Weekly reduction in housing benefit = $60/month ÷ 4.33 weeks/month ≈ $13.86/week.\n", "\n", "3. **Calculate the net weekly effect:**\n", " - Net weekly\n", @@ -5123,7 +5149,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Processed prompts: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 1/1 [00:16<00:00, 16.59s/it, est. speed input: 1.27 toks/s, output: 9.89 toks/s]\n" + "Processed prompts: 100%|██████████| 1/1 [00:16<00:00, 16.59s/it, est. speed input: 1.27 toks/s, output: 9.89 toks/s]\n" ] }, { @@ -5205,7 +5231,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Processed prompts: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 1/1 [00:27<00:00, 27.72s/it, est. speed input: 1.70 toks/s, output: 10.03 toks/s]\n" + "Processed prompts: 100%|██████████| 1/1 [00:27<00:00, 27.72s/it, est. speed input: 1.70 toks/s, output: 10.03 toks/s]\n" ] }, { @@ -5403,7 +5429,7 @@ " \n", " \n", "\n", - " Join Discord if you need help + \u2b50\ufe0f Star us on Github \u2b50\ufe0f\n", + " Join Discord if you need help + ⭐️ Star us on Github ⭐️\n", "\n" ] } @@ -5685,7 +5711,7 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_e43ad27d5d304d1ebf9b374016409a97", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_51948945111f437c9ed6ccab22072dd3", "value": "" } @@ -5706,7 +5732,7 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4c9248100f89400d9e1407dbb168d5d6", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_96cea0d773c8426b8be72dd7f72e5a82", "value": "" } @@ -5956,9 +5982,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_9eed940f3815428583b4ddefc1a81469", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_0b9230e976b34a9ea85978cf22857012", - "value": "Loading\u2007safetensors\u2007checkpoint\u2007shards:\u2007100%\u2007Completed\u2007|\u20072/2\u2007[00:51<00:00,\u200725.46s/it]\n" + "value": "Loading safetensors checkpoint shards: 100% Completed | 2/2 [00:51<00:00, 25.46s/it]\n" } }, "d034c840e7f74177a7b07a188d666b8d": { @@ -5977,9 +6003,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_129dc789722b43439574390bba63b36a", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_8991360910ef417db03499f76f5fe323", - "value": "Loading\u2007safetensors\u2007checkpoint\u2007shards:\u2007100%\u2007Completed\u2007|\u20072/2\u2007[00:47<00:00,\u200723.63s/it]\n" + "value": "Loading safetensors checkpoint shards: 100% Completed | 2/2 [00:47<00:00, 23.63s/it]\n" } }, "d6d5a7d96a034247b38d25d8a9cc979c": { @@ -6114,4 +6140,4 @@ }, "nbformat": 4, "nbformat_minor": 0 -} \ No newline at end of file +} diff --git a/nb/Qwen2.5_(3B)-GRPO.ipynb b/nb/Qwen2.5_(3B)-GRPO.ipynb index b10237a..150c0ed 100644 --- a/nb/Qwen2.5_(3B)-GRPO.ipynb +++ b/nb/Qwen2.5_(3B)-GRPO.ipynb @@ -8,7 +8,7 @@ "
\n", "\n", "\n", - " Join Discord if you need help + \u2b50 Star us on Github \u2b50\n", + " Join Discord if you need help + ⭐ Star us on Github ⭐\n", "
\n", "\n", "To install Unsloth on your own computer, follow the installation instructions on our Github page [here](https://docs.unsloth.ai/get-started/installing-+-updating).\n", @@ -48,14 +48,40 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\nimport os\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n # If you're not in Colab, just use pip install or uv pip install\n !pip install unsloth vllm\nelse:\n pass # For Colab / Kaggle, we need extra instructions hidden below \\/" + "source": [ + "%%capture\n", + "import os\n", + "if \"COLAB_\" not in \"\".join(os.environ.keys()):\n", + " # If you're not in Colab, just use pip install or uv pip install\n", + " !pip install unsloth vllm\n", + "else:\n", + " pass # For Colab / Kaggle, we need extra instructions hidden below \\/" + ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], - "source": "#@title Colab Extra Install { display-mode: \"form\" }\n%%capture\nimport os\n!pip install --upgrade -qqq uv\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n # If you're not in Colab, just use pip install!\n !pip install unsloth vllm\nelse:\n try: import numpy; get_numpy = f\"numpy=={numpy.__version__}\"\n except: get_numpy = \"numpy\"\n try: import subprocess; is_t4 = \"Tesla T4\" in str(subprocess.check_output([\"nvidia-smi\"]))\n except: is_t4 = False\n get_vllm, get_triton = (\"vllm==0.10.1\", \"triton==3.2.0\") if is_t4 else (\"vllm\", \"triton\")\n !uv pip install -qqq --upgrade \\\n unsloth {get_vllm} {get_numpy} torchvision bitsandbytes xformers\n !uv pip install -qqq {get_triton}\n!uv pip install transformers==4.55.4" + "source": [ + "#@title Colab Extra Install { display-mode: \"form\" }\n", + "%%capture\n", + "import os\n", + "!pip install --upgrade -qqq uv\n", + "if \"COLAB_\" not in \"\".join(os.environ.keys()):\n", + " # If you're not in Colab, just use pip install!\n", + " !pip install unsloth vllm\n", + "else:\n", + " try: import numpy; get_numpy = f\"numpy=={numpy.__version__}\"\n", + " except: get_numpy = \"numpy\"\n", + " try: import subprocess; is_t4 = \"Tesla T4\" in str(subprocess.check_output([\"nvidia-smi\"]))\n", + " except: is_t4 = False\n", + " get_vllm, get_triton = (\"vllm==0.10.1\", \"triton==3.2.0\") if is_t4 else (\"vllm\", \"triton\")\n", + " !uv pip install -qqq --upgrade \\\n", + " unsloth {get_vllm} {get_numpy} torchvision bitsandbytes xformers\n", + " !uv pip install -qqq {get_triton}\n", + "!uv pip install transformers==4.55.4" + ] }, { "cell_type": "markdown", @@ -452,7 +478,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Capturing CUDA graph shapes: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 27/27 [00:44<00:00, 1.66s/it]" + "Capturing CUDA graph shapes: 100%|██████████| 27/27 [00:44<00:00, 1.66s/it]" ] }, { @@ -832,14 +858,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", @@ -972,7 +998,7 @@ "\n", "Emily currently has 9 assignments with a total score of 92 points, and the final assignment she gets is 90 points. Therefore, her current total is:\n", "\\[ 9 \\times 92 + 90 = 828 + 90 = 918 \\]\n", - "She needs to find the minimum score to beat Ahmed, and her target would be greater than Adam\u2019s 813 points from his 9 assignments, so we calculate his total:\n", + "She needs to find the minimum score to beat Ahmed, and her target would be greater than Adam’s 813 points from his 9 assignments, so we calculate his total:\n", "\\[ 9 \\times 91 = 819 \\]\n", "To beat Emily, Ahmed needs a score such that his average score from all 10 \n", "Extracted:\n", @@ -981,7 +1007,7 @@ "\n", "Emily currently has 9 assignments with a total score of 92 points, and the final assignment she gets is 90 points. Therefore, her current total is:\n", "\\[ 9 \\times 92 + 90 = 828 + 90 = 918 \\]\n", - "She needs to find the minimum score to beat Ahmed, and her target would be greater than Adam\u2019s 813 points from his 9 assignments, so we calculate his total:\n", + "She needs to find the minimum score to beat Ahmed, and her target would be greater than Adam’s 813 points from his 9 assignments, so we calculate his total:\n", "\\[ 9 \\times 91 = 819 \\]\n", "To beat Emily, Ahmed needs a score such that his average score from all 10\n" ] @@ -3371,7 +3397,7 @@ "To find \\( x \\), we solve:\n", "\\[ 5 = \\frac{1}{\n", "-------------------- Question:\n", - "Clyde and Grace are building block towers. Grace\u2019s tower is 8 times the size of Clyde\u2019s at 40 inches tall. How many inches taller is Grace\u2019s tower than Clyde\u2019s? \n", + "Clyde and Grace are building block towers. Grace’s tower is 8 times the size of Clyde’s at 40 inches tall. How many inches taller is Grace’s tower than Clyde’s? \n", "Answer:\n", "35 \n", "Response:\n", @@ -3612,7 +3638,7 @@ "\n", "\\[ \\text{Initial Number of Bananas} -\n", "-------------------- Question:\n", - "At Theo\u2019s cafe, he makes 3 egg and 4 egg omelettes. His cafe is open from 7:00 a.m. to 11:00 a.m. In the first hour, 5 customers order the 3 egg omelettes. In the second hour, 7 customers order the 4 egg omelettes. In the third hour, 3 customers order the 3 egg omelettes. In the last hour, 8 customers order the 4 egg omelettes. How many eggs does Theo need to make all the omelettes? \n", + "At Theo’s cafe, he makes 3 egg and 4 egg omelettes. His cafe is open from 7:00 a.m. to 11:00 a.m. In the first hour, 5 customers order the 3 egg omelettes. In the second hour, 7 customers order the 4 egg omelettes. In the third hour, 3 customers order the 3 egg omelettes. In the last hour, 8 customers order the 4 egg omelettes. How many eggs does Theo need to make all the omelettes? \n", "Answer:\n", "84 \n", "Response:\n", @@ -3826,7 +3852,7 @@ "- First week: 40 new followers\n", "- Second week: Half of 40 (which is 20 followers)\n", "- Third week: Half of 20 (which is 10 followers)\n", - "- Total: 100 + 40 + 20 + 1\u6485c\n", + "- Total: 100 + 40 + 20 + 1撅c\n", "\n", "**Sarah's Followers:**\n", "- At the start: 50 followers\n", @@ -3843,7 +3869,7 @@ "- First week: 40 new followers\n", "- Second week: Half of 40 (which is 20 followers)\n", "- Third week: Half of 20 (which is 10 followers)\n", - "- Total: 100 + 40 + 20 + 1\u6485c\n", + "- Total: 100 + 40 + 20 + 1撅c\n", "\n", "**Sarah's Followers:**\n", "- At the start: 50 followers\n", @@ -3857,7 +3883,7 @@ "9 \n", "Response:\n", "\n", - "Jimmy starts with 18 cards. He gives 3 cards to Bob. That leaves Jimmy with 18 - 3 = 15 cards. Then he gives Mary twice as many cards as he gave to Bob. Giving Bob 3 cards, he gives Mary 2 * 3 = 6 cards. Hence, Jimmy\u2019s total number of cards is reduced by 6 more, so he is left with 15 - 6 = 9 cards.\n", + "Jimmy starts with 18 cards. He gives 3 cards to Bob. That leaves Jimmy with 18 - 3 = 15 cards. Then he gives Mary twice as many cards as he gave to Bob. Giving Bob 3 cards, he gives Mary 2 * 3 = 6 cards. Hence, Jimmy’s total number of cards is reduced by 6 more, so he is left with 15 - 6 = 9 cards.\n", "\n", "The final number of cards Jimmy has left is 9.\n", "\n", @@ -3975,7 +4001,7 @@ "\n", "To find the percentage of ads that are not interesting and don't get blocked, we need to break down the problem into several steps:\n", "\n", - "1. First, let\u2019s consider the total percentage of ads that are not blocked by AdBlock. According to the information, AdBlock blocks all but 20% of ads. This means that 80% of the ads get through AdBlock.\n", + "1. First, let’s consider the total percentage of ads that are not blocked by AdBlock. According to the information, AdBlock blocks all but 20% of ads. This means that 80% of the ads get through AdBlock.\n", "\n", "2. Of these 80% ads, 20% of them are interesting. Therefore, 80% of 80% = 80% * (80/100) = 64% are interested and pass through AdBlock.\n", "\n", @@ -3986,7 +4012,7 @@ "\n", "To find the percentage of ads that are not interesting and don't get blocked, we need to break down the problem into several steps:\n", "\n", - "1. First, let\u2019s consider the total percentage of ads that are not blocked by AdBlock. According to the information, AdBlock blocks all but 20% of ads. This means that 80% of the ads get through AdBlock.\n", + "1. First, let’s consider the total percentage of ads that are not blocked by AdBlock. According to the information, AdBlock blocks all but 20% of ads. This means that 80% of the ads get through AdBlock.\n", "\n", "2. Of these 80% ads, 20% of them are interesting. Therefore, 80% of 80% = 80% * (80/100) = 64% are interested and pass through AdBlock.\n", "\n", @@ -4061,7 +4087,7 @@ "\n", "To determine how many points Lizzie, Nathalie, and Aimee scored, we begin by calculating Lizzie's and Nathalie's scores. \n", "- Lizzie scored 4 points. \n", - "- Nathalie scored 3 more points than Lizzie, so Nathalie\u2019s score is:\n", + "- Nathalie scored 3 more points than Lizzie, so Nathalie’s score is:\n", " \\[\n", " 4 + 3 = 7 \\text{ points}\n", " \\]\n", @@ -4071,7 +4097,7 @@ " \\[\n", " 4 + 7 = 11 \\text{ points}\n", " \\]\n", - "- Aimee\u2019s score is twice this combined score:\n", + "- Aimee’s score is twice this combined score:\n", " \\[\n", " 2 \\times 11 = 22 \\text{ points}\n", " \\]\n", @@ -4083,7 +4109,7 @@ "\n", "To determine how many points Lizzie, Nathalie, and Aimee scored, we begin by calculating Lizzie's and Nathalie's scores. \n", "- Lizzie scored 4 points. \n", - "- Nathalie scored 3 more points than Lizzie, so Nathalie\u2019s score is:\n", + "- Nathalie scored 3 more points than Lizzie, so Nathalie’s score is:\n", " \\[\n", " 4 + 3 = 7 \\text{ points}\n", " \\]\n", @@ -4093,7 +4119,7 @@ " \\[\n", " 4 + 7 = 11 \\text{ points}\n", " \\]\n", - "- Aimee\u2019s score is twice this combined score:\n", + "- Aimee’s score is twice this combined score:\n", " \\[\n", " 2 \\times 11 = 22 \\text{ points}\n", " \\]\n", @@ -4363,9 +4389,9 @@ "3. Ken won 5 more games than Dave. Hence, the number of games Ken won is \\(10 + 5 = 15\\).\n", "\n", "Now, we can calculate the total number of games played by summing up the games won by all three:\n", - "- Ken\u2019s games: 15\n", - "- Dave\u2019s games: 10\n", - "- Jerry\u2019s games: 7\n", + "- Ken’s games: 15\n", + "- Dave’s games: 10\n", + "- Jerry’s games: 7\n", "\n", "The total number of games they played is:\n", "\\[ 15 + 10 + 7 = 32 \\]\n", @@ -4380,9 +4406,9 @@ "3. Ken won 5 more games than Dave. Hence, the number of games Ken won is \\(10 + 5 = 15\\).\n", "\n", "Now, we can calculate the total number of games played by summing up the games won by all three:\n", - "- Ken\u2019s games: 15\n", - "- Dave\u2019s games: 10\n", - "- Jerry\u2019s games: 7\n", + "- Ken’s games: 15\n", + "- Dave’s games: 10\n", + "- Jerry’s games: 7\n", "\n", "The total number of games they played is:\n", "\\[ 15 + 10 + 7 = 32 \\]\n", @@ -5139,10 +5165,10 @@ "1 \n", "Response:\n", "\n", - "Duke needs 17 more points to tie the record and ends up breaking the record by 5 points, meaning he scored a total of 257 + 5 = 262 points in the final game. With 5 free throws (1 point each) and 4 regular baskets (2 points each) he scored, this brings his points from these to 5*(1) + 4*(2) = 5 + 8 = 13 points. This means the points he scored beyond the free throws and regular baskets must have been two-pointers or three-pointers, and they totaled 262 - 13 = 249 points. Since normally Duke scores normally 2 three-pointers per game, let\u2019s assume he scored x three-pointers in the final game. Therefore, the total points from three-pointers would be 3*x, and he scored 13 + 249 = \n", + "Duke needs 17 more points to tie the record and ends up breaking the record by 5 points, meaning he scored a total of 257 + 5 = 262 points in the final game. With 5 free throws (1 point each) and 4 regular baskets (2 points each) he scored, this brings his points from these to 5*(1) + 4*(2) = 5 + 8 = 13 points. This means the points he scored beyond the free throws and regular baskets must have been two-pointers or three-pointers, and they totaled 262 - 13 = 249 points. Since normally Duke scores normally 2 three-pointers per game, let’s assume he scored x three-pointers in the final game. Therefore, the total points from three-pointers would be 3*x, and he scored 13 + 249 = \n", "Extracted:\n", "\n", - "Duke needs 17 more points to tie the record and ends up breaking the record by 5 points, meaning he scored a total of 257 + 5 = 262 points in the final game. With 5 free throws (1 point each) and 4 regular baskets (2 points each) he scored, this brings his points from these to 5*(1) + 4*(2) = 5 + 8 = 13 points. This means the points he scored beyond the free throws and regular baskets must have been two-pointers or three-pointers, and they totaled 262 - 13 = 249 points. Since normally Duke scores normally 2 three-pointers per game, let\u2019s assume he scored x three-pointers in the final game. Therefore, the total points from three-pointers would be 3*x, and he scored 13 + 249 =\n", + "Duke needs 17 more points to tie the record and ends up breaking the record by 5 points, meaning he scored a total of 257 + 5 = 262 points in the final game. With 5 free throws (1 point each) and 4 regular baskets (2 points each) he scored, this brings his points from these to 5*(1) + 4*(2) = 5 + 8 = 13 points. This means the points he scored beyond the free throws and regular baskets must have been two-pointers or three-pointers, and they totaled 262 - 13 = 249 points. Since normally Duke scores normally 2 three-pointers per game, let’s assume he scored x three-pointers in the final game. Therefore, the total points from three-pointers would be 3*x, and he scored 13 + 249 =\n", "-------------------- Question:\n", "Corey downloaded two movie series from his Netflix account with 12 and 14 seasons per series, respectively. However, in the week, his computer got a mechanical failure, and he lost two episodes from each season for both series. If each season in the movie series that Corey downloaded had 16 episodes, how many episodes remained after the computer's mechanical failure? \n", "Answer:\n", @@ -5324,7 +5350,7 @@ "- Cost for ten oranges: $46 \\times 10 = $460\n", "- Cost for ten mangoes: $57.50 \\times\n", "-------------------- Question:\n", - "The lights in Malcolm\u2019s house are flickering, and he hopes that replacing all of his white lights with colored lights will make it stop. He buys 12 red lights, 3 times as many blue lights, and 6 green lights. If he still has 5 colored lights left to buy, how many white lights did Malcolm have initially? \n", + "The lights in Malcolm’s house are flickering, and he hopes that replacing all of his white lights with colored lights will make it stop. He buys 12 red lights, 3 times as many blue lights, and 6 green lights. If he still has 5 colored lights left to buy, how many white lights did Malcolm have initially? \n", "Answer:\n", "59 \n", "Response:\n", @@ -6074,30 +6100,30 @@ "Extracted:\n", "$37500\n", "-------------------- Question:\n", - "36 liters of diesel fuel is worth \u20ac18. The tank of this pickup truck can hold 64 liters. How much does a full tank of diesel fuel cost? \n", + "36 liters of diesel fuel is worth €18. The tank of this pickup truck can hold 64 liters. How much does a full tank of diesel fuel cost? \n", "Answer:\n", "32 \n", "Response:\n", "\n", - "To determine the cost of a full tank of diesel fuel, we first need to identify the cost per liter of diesel fuel. Given that 36 liters of fuel cost \u20ac18, we can find the cost per liter by dividing the total cost by the number of liters:\n", + "To determine the cost of a full tank of diesel fuel, we first need to identify the cost per liter of diesel fuel. Given that 36 liters of fuel cost €18, we can find the cost per liter by dividing the total cost by the number of liters:\n", "\n", "\\[\n", - "\\text{Cost per liter} = \\frac{\\text{Total Cost}}{\\text{Total Liters}} = \\frac{18 \\, \\text{\u20ac}}{36 \\, \\text{L}} = 0.5 \\, \\text{\u20ac/L}\n", + "\\text{Cost per liter} = \\frac{\\text{Total Cost}}{\\text{Total Liters}} = \\frac{18 \\, \\text{€}}{36 \\, \\text{L}} = 0.5 \\, \\text{€/L}\n", "\\]\n", "\n", - "Now that we know the cost per liter is \u20ac0.5, we can find out how much 64 liters would cost by multiplying the cost per liter by the total number of liters the truck's tank can hold:\n", + "Now that we know the cost per liter is €0.5, we can find out how much 64 liters would cost by multiplying the cost per liter by the total number of liters the truck's tank can hold:\n", "\n", "\\[\n", "\\text{Cost for 64 liters} = 64 \\, \\text{L} \\times 0.5 \\, \n", "Extracted:\n", "\n", - "To determine the cost of a full tank of diesel fuel, we first need to identify the cost per liter of diesel fuel. Given that 36 liters of fuel cost \u20ac18, we can find the cost per liter by dividing the total cost by the number of liters:\n", + "To determine the cost of a full tank of diesel fuel, we first need to identify the cost per liter of diesel fuel. Given that 36 liters of fuel cost €18, we can find the cost per liter by dividing the total cost by the number of liters:\n", "\n", "\\[\n", - "\\text{Cost per liter} = \\frac{\\text{Total Cost}}{\\text{Total Liters}} = \\frac{18 \\, \\text{\u20ac}}{36 \\, \\text{L}} = 0.5 \\, \\text{\u20ac/L}\n", + "\\text{Cost per liter} = \\frac{\\text{Total Cost}}{\\text{Total Liters}} = \\frac{18 \\, \\text{€}}{36 \\, \\text{L}} = 0.5 \\, \\text{€/L}\n", "\\]\n", "\n", - "Now that we know the cost per liter is \u20ac0.5, we can find out how much 64 liters would cost by multiplying the cost per liter by the total number of liters the truck's tank can hold:\n", + "Now that we know the cost per liter is €0.5, we can find out how much 64 liters would cost by multiplying the cost per liter by the total number of liters the truck's tank can hold:\n", "\n", "\\[\n", "\\text{Cost for 64 liters} = 64 \\, \\text{L} \\times 0.5 \\,\n", @@ -6248,7 +6274,7 @@ "\n", "To determine the total value of Smaug's hoard expressed as a number of copper coins, we need to consider the value of each type of coin in terms of copper coins, and then sum them up.\n", "\n", - "- First, let\u2019s convert the silver coins to copper coins. Given each silver coin is worth 8 copper coins, 60 silver coins will be worth:\n", + "- First, let’s convert the silver coins to copper coins. Given each silver coin is worth 8 copper coins, 60 silver coins will be worth:\n", " \\[\n", " 60 \\text{ silver coins} \\times 8 \\text{ copper coins per silver coin} = 480 \\text{ copper coins}\n", " \\]\n", @@ -6263,7 +6289,7 @@ "\n", "To determine the total value of Smaug's hoard expressed as a number of copper coins, we need to consider the value of each type of coin in terms of copper coins, and then sum them up.\n", "\n", - "- First, let\u2019s convert the silver coins to copper coins. Given each silver coin is worth 8 copper coins, 60 silver coins will be worth:\n", + "- First, let’s convert the silver coins to copper coins. Given each silver coin is worth 8 copper coins, 60 silver coins will be worth:\n", " \\[\n", " 60 \\text{ silver coins} \\times 8 \\text{ copper coins per silver coin} = 480 \\text{ copper coins}\n", " \\]\n", @@ -6385,7 +6411,7 @@ "Extracted:\n", "54\n", - "Tate initially has 32 tickets. He buys 2 more tickets, so his total number of tickets becomes 32 + 2 = 34. We need to determine Peyton\u2019s number of tickets. His friend Peyton has half as many tickets as Tate. So, Peyton has 34 / 2 = 17 tickets. Together, Tate and Peyton have 34 + 17 = 51 tickets.\n", + "Tate initially has 32 tickets. He buys 2 more tickets, so his total number of tickets becomes 32 + 2 = 34. We need to determine Peyton’s number of tickets. His friend Peyton has half as many tickets as Tate. So, Peyton has 34 / 2 = 17 tickets. Together, Tate and Peyton have 34 + 17 = 51 tickets.\n", "\n", "51 \n", "Extracted:\n", @@ -6645,7 +6671,7 @@ "15 \n", "Response:\n", "\n", - "The total weight of swords that need to be transported is 1200 pounds. There are 10 squads and each squad consists of 8 orcs. Therefore, the total number of orcs is 10 * 8 = 80 orcs. To find out how many pounds of swords each orc needs to carry, divide the total weight by the total number of orcs. Thus, 1200 pounds \u00f7 80 orcs = 15 pounds per orc. \n", + "The total weight of swords that need to be transported is 1200 pounds. There are 10 squads and each squad consists of 8 orcs. Therefore, the total number of orcs is 10 * 8 = 80 orcs. To find out how many pounds of swords each orc needs to carry, divide the total weight by the total number of orcs. Thus, 1200 pounds ÷ 80 orcs = 15 pounds per orc. \n", "\n", "15 \n", "Extracted:\n", @@ -6662,7 +6688,7 @@ "\n", "Since each round around the island takes 30 minutes, calculating the total time spent is straightforward. It requires multiplying the total number of rounds by the time it takes to complete one rotation.\n", "\n", - "Total time = 26 rounds \u00d7 30 minutes per round = 780 minutes.\n", + "Total time = 26 rounds × 30 minutes per round = 780 minutes.\n", "\n", "Thus, the Polar Bears spent 780 minutes circling the island over that weekend.\n", "\n", @@ -6971,8 +6997,8 @@ "\n", "Kenzo originally has 80 office chairs with 5 legs each and 20 round tables with 3 legs each. So, the total number of legs from the chairs and tables is calculated as follows: \n", "\n", - "Number of legs from chairs = 80 chairs \u00d7 5 legs per chair = 400 legs\n", - "Number of legs from tables = 20 tables \u00d7 3 legs per table = 60 legs\n", + "Number of legs from chairs = 80 chairs × 5 legs per chair = 400 legs\n", + "Number of legs from tables = 20 tables × 3 legs per table = 60 legs\n", "\n", "So, the total number of legs initially is: 400 legs + 60 legs = 460 legs\n", "\n", @@ -6983,8 +7009,8 @@ "\n", "Kenzo originally has 80 office chairs with 5 legs each and 20 round tables with 3 legs each. So, the total number of legs from the chairs and tables is calculated as follows: \n", "\n", - "Number of legs from chairs = 80 chairs \u00d7 5 legs per chair = 400 legs\n", - "Number of legs from tables = 20 tables \u00d7 3 legs per table = 60 legs\n", + "Number of legs from chairs = 80 chairs × 5 legs per chair = 400 legs\n", + "Number of legs from tables = 20 tables × 3 legs per table = 60 legs\n", "\n", "So, the total number of legs initially is: 400 legs + 60 legs = 460 legs\n", "\n", @@ -7020,7 +7046,7 @@ "\n", "The total number of people the bus carried\n", "-------------------- Question:\n", - "Sharon\u2019s vacation rental has a Keurig coffee machine. She will be there for 40 days. She has 3 cups of coffee (3 coffee pods) every morning. Her coffee pods come 30 pods to a box for $8.00. How much will she spend on coffee to last her for the entire vacation? \n", + "Sharon’s vacation rental has a Keurig coffee machine. She will be there for 40 days. She has 3 cups of coffee (3 coffee pods) every morning. Her coffee pods come 30 pods to a box for $8.00. How much will she spend on coffee to last her for the entire vacation? \n", "Answer:\n", "32 \n", "Response:\n", @@ -7066,7 +7092,7 @@ "12 \n", "Response:\n", "\n", - "To find out how many of the computers sold are desktops, we first calculate the number of laptops and netbooks sold based on the given fractions. Mr. Lu\u2019s store sold a total of 72 computers. \n", + "To find out how many of the computers sold are desktops, we first calculate the number of laptops and netbooks sold based on the given fractions. Mr. Lu’s store sold a total of 72 computers. \n", "\n", "Since half of their sales are laptops, we can calculate the number of laptops sold as:\n", "\\[ \\text{Number of laptops} = \\frac{1}{2} \\times 72 = 36 \\]\n", @@ -7078,7 +7104,7 @@ "\\[ \\text{Number of desktops} = 72 - (\\text{Number of laptops} + \\text{Number of netbooks \n", "Extracted:\n", "\n", - "To find out how many of the computers sold are desktops, we first calculate the number of laptops and netbooks sold based on the given fractions. Mr. Lu\u2019s store sold a total of 72 computers. \n", + "To find out how many of the computers sold are desktops, we first calculate the number of laptops and netbooks sold based on the given fractions. Mr. Lu’s store sold a total of 72 computers. \n", "\n", "Since half of their sales are laptops, we can calculate the number of laptops sold as:\n", "\\[ \\text{Number of laptops} = \\frac{1}{2} \\times 72 = 36 \\]\n", @@ -7257,7 +7283,7 @@ "12 \n", "Response:\n", "\n", - "According to the problem, Ken's house is twice as far from Dawn's house as Mary's house is from Dawn's house. We are given that Ken's house is 4 miles away from Dawn's house. Given this information, we can deduce that Mary's house is 2 miles away from Dawn\u2019s house (since Ken\u2019s house is twice as far from Dawn's house as Mary's house is).\n", + "According to the problem, Ken's house is twice as far from Dawn's house as Mary's house is from Dawn's house. We are given that Ken's house is 4 miles away from Dawn's house. Given this information, we can deduce that Mary's house is 2 miles away from Dawn’s house (since Ken’s house is twice as far from Dawn's house as Mary's house is).\n", "\n", "First, we calculate the distance Ken travels from his house to Dawn's house:\n", "\\[ 4 \\text{ miles} \\]\n", @@ -7265,13 +7291,13 @@ "Next, Ken travels from Dawn's house to Mary's house, which is 2 miles:\n", "\\[ 2 \\text{ miles} \\]\n", "\n", - "Then, Ken travels back from Mary\u2019s house to Dawn\u2019s house, which is 2 miles:\n", + "Then, Ken travels back from Mary’s house to Dawn’s house, which is 2 miles:\n", "\\[ 2 \\text{ miles} \\]\n", "\n", - "Finally, Ken travels from Dawn\u2019s house back to his own house. By the problem statement, Ken's house is 4 miles away from \n", + "Finally, Ken travels from Dawn’s house back to his own house. By the problem statement, Ken's house is 4 miles away from \n", "Extracted:\n", "\n", - "According to the problem, Ken's house is twice as far from Dawn's house as Mary's house is from Dawn's house. We are given that Ken's house is 4 miles away from Dawn's house. Given this information, we can deduce that Mary's house is 2 miles away from Dawn\u2019s house (since Ken\u2019s house is twice as far from Dawn's house as Mary's house is).\n", + "According to the problem, Ken's house is twice as far from Dawn's house as Mary's house is from Dawn's house. We are given that Ken's house is 4 miles away from Dawn's house. Given this information, we can deduce that Mary's house is 2 miles away from Dawn’s house (since Ken’s house is twice as far from Dawn's house as Mary's house is).\n", "\n", "First, we calculate the distance Ken travels from his house to Dawn's house:\n", "\\[ 4 \\text{ miles} \\]\n", @@ -7279,10 +7305,10 @@ "Next, Ken travels from Dawn's house to Mary's house, which is 2 miles:\n", "\\[ 2 \\text{ miles} \\]\n", "\n", - "Then, Ken travels back from Mary\u2019s house to Dawn\u2019s house, which is 2 miles:\n", + "Then, Ken travels back from Mary’s house to Dawn’s house, which is 2 miles:\n", "\\[ 2 \\text{ miles} \\]\n", "\n", - "Finally, Ken travels from Dawn\u2019s house back to his own house. By the problem statement, Ken's house is 4 miles away from\n", + "Finally, Ken travels from Dawn’s house back to his own house. By the problem statement, Ken's house is 4 miles away from\n", "-------------------- Question:\n", "Wilson goes to a fast-food restaurant. He buys 2 hamburgers for $5 each and 3 bottles of cola for $2 each. Wilson uses his $4 discount coupon. How much money does he pay in total? \n", "Answer:\n", @@ -7290,9 +7316,9 @@ "Response:\n", "\n", "Wilson buys 2 hamburgers at $5 each, so the total for hamburgers is:\n", - "2 \u00d7 $5 = $10\n", + "2 × $5 = $10\n", "He buys 3 bottles of cola at $2 each, so the total for cola is:\n", - "3 \u00d7 $2 = $6\n", + "3 × $2 = $6\n", "Adding both the total amount for hamburgers and cola gives us:\n", "$10 + $6 = $16\n", "After using his $4 discount coupon, the total amount he has to pay is:\n", @@ -7476,7 +7502,7 @@ "Extracted:\n", "230\n", "-------------------- Question:\n", - "Michael\u2019s largest watermelon weighs 8 pounds. His neighbor, Clay, grew a watermelon three times that size. And their mutual friend, John, grew a watermelon that was half the size of Clay\u2019s. How big was John\u2019s watermelon? \n", + "Michael’s largest watermelon weighs 8 pounds. His neighbor, Clay, grew a watermelon three times that size. And their mutual friend, John, grew a watermelon that was half the size of Clay’s. How big was John’s watermelon? \n", "Answer:\n", "12 \n", "Response:\n", @@ -7844,7 +7870,7 @@ "1. **Possible combinations with pepperoni and without peppers:**\n", " - Meat: 4 (excluding pepperoni)\n", " - Vegetable: 4 (excluding peppers)\n", - " - Total combinations with this setup = 4 (meat) \u00d7 4 (vegetable) = 16\n", + " - Total combinations with this setup = 4 (meat) × 4 (vegetable) = 16\n", "\n", "2. **Possible combinations without pepperoni:**\n", " \n", @@ -7859,7 +7885,7 @@ "1. **Possible combinations with pepperoni and without peppers:**\n", " - Meat: 4 (excluding pepperoni)\n", " - Vegetable: 4 (excluding peppers)\n", - " - Total combinations with this setup = 4 (meat) \u00d7 4 (vegetable) = 16\n", + " - Total combinations with this setup = 4 (meat) × 4 (vegetable) = 16\n", "\n", "2. **Possible combinations without pepperoni:**\n", "-------------------- Question:\n", @@ -7967,7 +7993,7 @@ "12 \n", "Response:\n", "\n", - "The total number of people at the reunion was the sum of kids and adults, which is 45 kids + 123 adults = 168 people. Since there are 14 tables, to find out how many people sat at each table, we divide the total number of people by the number of tables: 168 people \u00f7 14 tables = 12 people per table.\n", + "The total number of people at the reunion was the sum of kids and adults, which is 45 kids + 123 adults = 168 people. Since there are 14 tables, to find out how many people sat at each table, we divide the total number of people by the number of tables: 168 people ÷ 14 tables = 12 people per table.\n", "\n", "12 \n", "Extracted:\n", @@ -8079,7 +8105,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Processed prompts: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 1/1 [00:00<00:00, 1.71it/s, est. speed input: 63.38 toks/s, output: 25.69 toks/s]\n" + "Processed prompts: 100%|██████████| 1/1 [00:00<00:00, 1.71it/s, est. speed input: 63.38 toks/s, output: 25.69 toks/s]\n" ] }, { @@ -8161,7 +8187,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Processed prompts: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 1/1 [00:03<00:00, 3.06s/it, est. speed input: 14.05 toks/s, output: 29.09 toks/s]\n" + "Processed prompts: 100%|██████████| 1/1 [00:03<00:00, 3.06s/it, est. speed input: 14.05 toks/s, output: 29.09 toks/s]\n" ] }, { @@ -8313,7 +8339,7 @@ " \n", " \n", "\n", - " Join Discord if you need help + \u2b50\ufe0f Star us on Github \u2b50\ufe0f\n", + " Join Discord if you need help + ⭐️ Star us on Github ⭐️\n", "\n" ] } @@ -8680,9 +8706,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_9258ed2d94924de3b944105e280fbf5f", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_1c68fbbc833d4c3fbbb686f15b9ba939", - "value": "special_tokens_map.json:\u2007100%" + "value": "special_tokens_map.json: 100%" } }, "0cc22299405f4361b1b5f6234f35bae3": { @@ -8868,9 +8894,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_d6bd8a9af6e84500956bb565f2fe2c4e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_6643bd32eb984a929159c8d441530aaf", - "value": "\u2007605/605\u2007[00:00<00:00,\u200752.3kB/s]" + "value": " 605/605 [00:00<00:00, 52.3kB/s]" } }, "0f9dcd9b698b4af699cfafdf2874612c": { @@ -8889,9 +8915,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_f8198301b2db406685189589e59efd9a", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e813e26451e3476a82a9b75668b638a1", - "value": "\u20072.36G/2.36G\u2007[00:17<00:00,\u2007511MB/s]" + "value": " 2.36G/2.36G [00:17<00:00, 511MB/s]" } }, "1060550e9dec4fa59e2b9b488af61b48": { @@ -9081,9 +9107,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_6a49aa5a2b3646d4b772e2398eacf72d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_edeb0f283b0b495eab6e06a15fa38af0", - "value": "added_tokens.json:\u2007100%" + "value": "added_tokens.json: 100%" } }, "16e05d45084c44b68697316bc8791b88": { @@ -9126,9 +9152,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_989519f3e61b4132b3f88fbdf29703a9", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e0f87f6fd7c84261818624defe575a1c", - "value": "Map:\u2007100%" + "value": "Map: 100%" } }, "1793da35014840b08d776b4abc9a1212": { @@ -9147,9 +9173,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_231be59acb2a4921ba53ba97bdf5888a", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_5aa0e699560c4fe89250b4ccac6401cb", - "value": "train-00000-of-00001.parquet:\u2007100%" + "value": "train-00000-of-00001.parquet: 100%" } }, "1a619a0a53c149c3b9b0cbd53f1d2034": { @@ -9168,9 +9194,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c626fa7f5c624ec5a414e6f58d71ea96", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_2a8fb4b980ab45be81ae39d5c6b7bad4", - "value": "\u20077.94k/7.94k\u2007[00:00<00:00,\u2007445kB/s]" + "value": " 7.94k/7.94k [00:00<00:00, 445kB/s]" } }, "1b1e49195c5b48a1b6e4364db01c5c87": { @@ -9934,9 +9960,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_784a5c0d201d46459a160e0dc8857fd9", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_d753c1ecc6034237ba4401a007451bfb", - "value": "\u20071.67M/1.67M\u2007[00:00<00:00,\u20071.96MB/s]" + "value": " 1.67M/1.67M [00:00<00:00, 1.96MB/s]" } }, "2f9a64ad318248bf99fccdedafaddfc0": { @@ -9955,9 +9981,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_7c67cf4d83e34cd6a475405be45f345e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_c184e417ad134e94995a55d4bc7c94b5", - "value": "Generating\u2007train\u2007split:\u2007100%" + "value": "Generating train split: 100%" } }, "32719fe47c8543f69bcddc9b9bc0f719": { @@ -10022,9 +10048,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_5cf88ae3c2fc4419a71eb45f28f67384", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_327fa24e70ee4cae903fe5673bcf3bfc", - "value": "vocab.json:\u2007100%" + "value": "vocab.json: 100%" } }, "32defd049d8a473c8ad8fc954c0e3d13": { @@ -10043,9 +10069,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_694494752ab2458ead83b0331af20122", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_1cc89fa4d7c744c69c3be1a3d0b3ba2d", - "value": "tokenizer_config.json:\u2007100%" + "value": "tokenizer_config.json: 100%" } }, "32f42609c389450bbaad10f0a884c527": { @@ -10132,7 +10158,7 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c314f7a0c6084790a2620ec231297525", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_950c1b85a8bd456ab928df4cc3710b21", "value": "" } @@ -10530,9 +10556,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_b245d8e4c115438583ef8b433d4d16be", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_df7f9d244f70440aaf401c6a856a6ed0", - "value": "merges.txt:\u2007100%" + "value": "merges.txt: 100%" } }, "4379f29ff4a74df784bb2fb3989c64cc": { @@ -10551,9 +10577,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_62cbedb2d67d4a2c8888b46e56e39f65", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_57f2c072b0574f74a7dbe2a9d303cbd4", - "value": "\u20071.67M/1.67M\u2007[00:00<00:00,\u20072.32MB/s]" + "value": " 1.67M/1.67M [00:00<00:00, 2.32MB/s]" } }, "4382289897b54e63b22541d0b45c9bf3": { @@ -10612,9 +10638,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c6b2d407256549a5bc7543cb6d7475b4", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_545251749b10408d82c9a60b914dce34", - "value": "special_tokens_map.json:\u2007100%" + "value": "special_tokens_map.json: 100%" } }, "44b6ecc9b9fc436e92701ec2409e0dd1": { @@ -10633,9 +10659,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_96519710bf5047239f6800100471ee6e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e9b020ce2dda442b871b41e417210b52", - "value": "\u20071319/1319\u2007[00:00<00:00,\u200747778.62\u2007examples/s]" + "value": " 1319/1319 [00:00<00:00, 47778.62 examples/s]" } }, "451d181b9ef24b0e9fc3914a4466b9f1": { @@ -11033,9 +11059,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_e14a085ca7e6410ca78fe0f27533be6d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_6d204a683e2d49deb5f504f91ee72044", - "value": "tokenizer_config.json:\u2007100%" + "value": "tokenizer_config.json: 100%" } }, "589a3857411f49d4ae019465f536da45": { @@ -11210,9 +11236,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_11bd7b89fa794a899db9c83cb2a99df5", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_0e9d3d2244a54eb781311c01965c33e4", - "value": "\u20077473/7473\u2007[00:00<00:00,\u20078101.99\u2007examples/s]" + "value": " 7473/7473 [00:00<00:00, 8101.99 examples/s]" } }, "5cda6bd66e3248c4be477f1c1b40198d": { @@ -12014,9 +12040,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_78774135f2d1468d9172fa7588afef20", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_9b3233c52ecc4b259deed2d5d6f332eb", - "value": "Loading\u2007safetensors\u2007checkpoint\u2007shards:\u2007100%\u2007Completed\u2007|\u20071/1\u2007[00:01<00:00,\u2007\u20071.41s/it]\n" + "value": "Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:01<00:00,  1.41s/it]\n" } }, "7529f4d84b9f46709ec5f0e22b75baf4": { @@ -12087,9 +12113,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_269b131121554544a21d47596a24aa58", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_73758a3ca9c14d7eaa6504d8b9b6e333", - "value": "added_tokens.json:\u2007100%" + "value": "added_tokens.json: 100%" } }, "78345adf6d8143e3a649a05c5d476115": { @@ -12611,9 +12637,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_b131891a909743f5826429a84a74b4f9", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_3745da58146d445abe863a248e86d9be", - "value": "README.md:\u2007100%" + "value": "README.md: 100%" } }, "8a89212807f243d4b135a3188f4fb6d0": { @@ -12632,9 +12658,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c9b781db7ee2422b8dc4eb6aa2575d87", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_5100eb13df1c453980942ec78d0e810b", - "value": "\u200711.4M/11.4M\u2007[00:00<00:00,\u200740.1MB/s]" + "value": " 11.4M/11.4M [00:00<00:00, 40.1MB/s]" } }, "8aa00f7a2de64af8b0a5b8cd5fefb14e": { @@ -12691,9 +12717,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_52ba857bdfc742a2aae9e9a949054e5c", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_589a3857411f49d4ae019465f536da45", - "value": "\u20077.36k/7.36k\u2007[00:00<00:00,\u2007512kB/s]" + "value": " 7.36k/7.36k [00:00<00:00, 512kB/s]" } }, "8ee15a3d02954d31899d0b6a8dbf7988": { @@ -12809,9 +12835,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1ccb12282661417383a7f96bd5ea4eac", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_32a16b0cbbe548f4869cf665d737b9ba", - "value": "model.safetensors:\u2007100%" + "value": "model.safetensors: 100%" } }, "950c1b85a8bd456ab928df4cc3710b21": { @@ -13017,9 +13043,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_3b6d4b41992b43acac5daa31f119cd30", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_c5b4960fc3ba4459a6eaff43772050a3", - "value": "\u2007605/605\u2007[00:00<00:00,\u200748.4kB/s]" + "value": " 605/605 [00:00<00:00, 48.4kB/s]" } }, "989519f3e61b4132b3f88fbdf29703a9": { @@ -13090,9 +13116,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a422e26291bb47d08e36b7d94fffcc9b", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_799dc8fe6f7d4e4db68ebd6b1d8fdbd5", - "value": "vocab.json:\u2007100%" + "value": "vocab.json: 100%" } }, "9929aaffe7614506ae934a14d2b5c3ca": { @@ -13178,9 +13204,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_29a3f867c8bd4769b8e649b8912776a2", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_2317c9a94dfe4a62a8605038ebc2210a", - "value": "\u2007271/271\u2007[00:00<00:00,\u200715.4kB/s]" + "value": " 271/271 [00:00<00:00, 15.4kB/s]" } }, "9bcb0000da6e434d9b241a09d9b46244": { @@ -13301,9 +13327,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_420154f0bbe344a7979769a8ef066df7", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_1ee34b9171604805887aee3baa4ff099", - "value": "\u20072.31M/2.31M\u2007[00:00<00:00,\u200725.1MB/s]" + "value": " 2.31M/2.31M [00:00<00:00, 25.1MB/s]" } }, "a422e26291bb47d08e36b7d94fffcc9b": { @@ -13450,9 +13476,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_5d190801b10b49a698c3a69c7cd1f253", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_0112cfe0a27e43fd9509ef8da050e188", - "value": "\u2007614/614\u2007[00:00<00:00,\u200752.1kB/s]" + "value": " 614/614 [00:00<00:00, 52.1kB/s]" } }, "a8501b8551d4431fbb708d6556e7dd85": { @@ -13590,7 +13616,7 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_24757877452d48f0afe92bb66e2c48ac", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_b64be6a0d6df44899ba6fe0a47f676e1", "value": "" } @@ -13837,9 +13863,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_37e2f38471d4481eb93031ca52bdb785", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_067aaf24d87a400b9f636762051598f3", - "value": "Loading\u2007safetensors\u2007checkpoint\u2007shards:\u2007100%\u2007Completed\u2007|\u20071/1\u2007[00:02<00:00,\u2007\u20072.58s/it]\n" + "value": "Loading safetensors checkpoint shards: 100% Completed | 1/1 [00:02<00:00,  2.58s/it]\n" } }, "b64be6a0d6df44899ba6fe0a47f676e1": { @@ -13873,9 +13899,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4ddc8fa0c9834cd3b36274a9e0e05504", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_ec13219478284bbdbe69cbcd3dff5c34", - "value": "\u2007614/614\u2007[00:00<00:00,\u200741.6kB/s]" + "value": " 614/614 [00:00<00:00, 41.6kB/s]" } }, "bc81dfa869f94f059545a7d59b7eb70a": { @@ -13931,9 +13957,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_71898e427b1245d682178a90c6752e8e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_90295f1308054c66b557da231d25d3d5", - "value": "\u2007419k/419k\u2007[00:00<00:00,\u200721.7MB/s]" + "value": " 419k/419k [00:00<00:00, 21.7MB/s]" } }, "c13ff304d7bc4c67a1f51da5d2303f96": { @@ -14385,9 +14411,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c404db8298364742b8ce076cddcb7641", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_03aca6d92597461481419eac9bdb6635", - "value": "tokenizer.json:\u2007100%" + "value": "tokenizer.json: 100%" } }, "c9b781db7ee2422b8dc4eb6aa2575d87": { @@ -14510,9 +14536,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_95fa6a54e7e543c7b2dc5d5cf04daf58", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_8ee15a3d02954d31899d0b6a8dbf7988", - "value": "Generating\u2007test\u2007split:\u2007100%" + "value": "Generating test split: 100%" } }, "d13cc9053e7a4982b6fc7f00474f25c7": { @@ -14789,9 +14815,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_d36027362b41413fa3ee4670dd820a88", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_d94e3315b99747c39190fee9b79f00ac", - "value": "\u20077473/7473\u2007[00:00<00:00,\u200716342.30\u2007examples/s]" + "value": " 7473/7473 [00:00<00:00, 16342.30 examples/s]" } }, "df7f9d244f70440aaf401c6a856a6ed0": { @@ -14954,9 +14980,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_aaf79572906c4443b5542bef5dee8055", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_bc81dfa869f94f059545a7d59b7eb70a", - "value": "generation_config.json:\u2007100%" + "value": "generation_config.json: 100%" } }, "e505a51d928542b883743a6d48c5cbd3": { @@ -15253,9 +15279,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1f88a409f3c64ea291861837e830ce39", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e9881edb2c7846fea2070c8e724b25cb", - "value": "\u200711.4M/11.4M\u2007[00:00<00:00,\u200742.5MB/s]" + "value": " 11.4M/11.4M [00:00<00:00, 42.5MB/s]" } }, "edeb0f283b0b495eab6e06a15fa38af0": { @@ -15313,9 +15339,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_0c2e39b20d5941969101b793b94fdbfa", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_f0c74de390814c2b9f38f7cc02458426", - "value": "test-00000-of-00001.parquet:\u2007100%" + "value": "test-00000-of-00001.parquet: 100%" } }, "f0c74de390814c2b9f38f7cc02458426": { @@ -15477,9 +15503,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_9929aaffe7614506ae934a14d2b5c3ca", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_1060550e9dec4fa59e2b9b488af61b48", - "value": "\u20072.78M/2.78M\u2007[00:00<00:00,\u20073.27MB/s]" + "value": " 2.78M/2.78M [00:00<00:00, 3.27MB/s]" } }, "f77e9e785cc04c00822fc8b7ef2959da": { @@ -15618,9 +15644,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_32f42609c389450bbaad10f0a884c527", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_a9e874a076234974a102b02551ec8666", - "value": "\u20077.36k/7.36k\u2007[00:00<00:00,\u2007580kB/s]" + "value": " 7.36k/7.36k [00:00<00:00, 580kB/s]" } }, "f89bfff6526c43bcb75c93f987bfba72": { @@ -15639,9 +15665,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c69730dfeec14cedaf8917ee9ad27dcb", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_00177cac24374c7d9144e3856594ddaf", - "value": "tokenizer.json:\u2007100%" + "value": "tokenizer.json: 100%" } }, "fa996ce2f8f24a5d80914c4ef5b6e788": { @@ -15712,9 +15738,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_878702a019e34789b5b2f68eb321f2c8", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_91998be216844c079144c18278f660d7", - "value": "\u20072.78M/2.78M\u2007[00:01<00:00,\u20072.68MB/s]" + "value": " 2.78M/2.78M [00:01<00:00, 2.68MB/s]" } }, "fc87fbccba304554be90701b39429918": { @@ -15785,9 +15811,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_0e8c4086734c4296884db9c5b67ddd33", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_c3a97f3130bc42c1be2f8c7dc81ca6cc", - "value": "merges.txt:\u2007100%" + "value": "merges.txt: 100%" } }, "fdcbb976f35e4aa383d9ebcd7304849f": { @@ -15864,4 +15890,4 @@ }, "nbformat": 4, "nbformat_minor": 0 -} \ No newline at end of file +} diff --git a/original_template/Llama3.1_(8B)-GRPO.ipynb b/original_template/Llama3.1_(8B)-GRPO.ipynb index 006d7b2..e98d943 100644 --- a/original_template/Llama3.1_(8B)-GRPO.ipynb +++ b/original_template/Llama3.1_(8B)-GRPO.ipynb @@ -445,7 +445,7 @@ }, { "cell_type": "code", - "execution_count": 3, + "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/", @@ -673,14 +673,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", diff --git a/original_template/Mistral_v0.3_(7B)-GRPO.ipynb b/original_template/Mistral_v0.3_(7B)-GRPO.ipynb index 058abbf..17865f8 100644 --- a/original_template/Mistral_v0.3_(7B)-GRPO.ipynb +++ b/original_template/Mistral_v0.3_(7B)-GRPO.ipynb @@ -525,14 +525,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", diff --git a/original_template/Phi_4_(14B)-GRPO.ipynb b/original_template/Phi_4_(14B)-GRPO.ipynb index 2aea67b..4b49e8e 100644 --- a/original_template/Phi_4_(14B)-GRPO.ipynb +++ b/original_template/Phi_4_(14B)-GRPO.ipynb @@ -297,14 +297,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n", diff --git a/original_template/Qwen2.5_(3B)-GRPO.ipynb b/original_template/Qwen2.5_(3B)-GRPO.ipynb index f671a3d..9a4262c 100644 --- a/original_template/Qwen2.5_(3B)-GRPO.ipynb +++ b/original_template/Qwen2.5_(3B)-GRPO.ipynb @@ -814,14 +814,14 @@ " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\"^\\n.*?\\n\\n\\n.*?\\n\\n$\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def soft_format_reward_func(completions, **kwargs) -> list[float]:\n", " \"\"\"Reward function that checks if the completion has a specific format.\"\"\"\n", " pattern = r\".*?\\s*.*?\"\n", " responses = [completion[0][\"content\"] for completion in completions]\n", - " matches = [re.match(pattern, r) for r in responses]\n", + " matches = [re.match(pattern, r, flags=re.DOTALL) for r in responses]\n", " return [0.5 if match else 0.0 for match in matches]\n", "\n", "def count_xml(text) -> float:\n",