|
| 1 | +<div style="background: linear-gradient(135deg, #2563eb, #1e40af); padding: 24px; border-radius: 8px; color: white; text-align: center; margin-bottom: 24px;"> |
| 2 | + <h2 style="margin: 0; color: white;">🚀 GPT-OSS Open Source Language Models</h2> |
| 3 | + <p style="margin: 8px 0 0 0; opacity: 0.9;">High-performance lightweight AI models with open innovation under Apache 2.0 license</p> |
| 4 | +</div> |
| 5 | + |
| 6 | +## 🎯 Core Highlights |
| 7 | + |
| 8 | +<div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(300px, 1fr)); gap: 16px; margin: 16px 0;"> |
| 9 | + |
| 10 | +<div style="background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 8px; padding: 20px;"> |
| 11 | +<h3 style="margin-top: 0; color: #1e40af;">🏆 Exceptional Performance</h3> |
| 12 | +<ul style="margin: 0; padding-left: 20px;"> |
| 13 | + <li><strong>GPT-OSS-120B</strong>: Matches OpenAI o4-mini performance</li> |
| 14 | + <li><strong>GPT-OSS-20B</strong>: Achieves OpenAI o3-mini level</li> |
| 15 | + <li><strong>Reasoning Advantage</strong>: Best performance among open models of similar scale</li> |
| 16 | + <li><strong>Tool Integration</strong>: Powerful tool usage capabilities</li> |
| 17 | +</ul> |
| 18 | +</div> |
| 19 | + |
| 20 | +<div style="background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 8px; padding: 20px;"> |
| 21 | +<h3 style="margin-top: 0; color: #1e40af;">💻 Hardware Friendly</h3> |
| 22 | +<ul style="margin: 0; padding-left: 20px;"> |
| 23 | + <li><strong>120B Model</strong>: Runs on a single 80GB GPU</li> |
| 24 | + <li><strong>20B Model</strong>: Requires only 16GB memory edge devices</li> |
| 25 | + <li><strong>Consumer Hardware</strong>: Optimized deployment costs</li> |
| 26 | + <li><strong>Edge Computing</strong>: Local inference without cloud dependency</li> |
| 27 | +</ul> |
| 28 | +</div> |
| 29 | + |
| 30 | +</div> |
| 31 | + |
| 32 | +## 🔬 Technical Features |
| 33 | + |
| 34 | +<div style="background: #eff6ff; border-left: 4px solid #2563eb; padding: 16px; margin: 16px 0; border-radius: 4px;"> |
| 35 | +<strong>🧠 Advanced Training Technology</strong><br> |
| 36 | +Trained using reinforcement learning combined with techniques inspired by OpenAI's most advanced internal models (including o3 and other cutting-edge systems), ensuring exceptional performance in reasoning and tool usage. |
| 37 | +</div> |
| 38 | + |
| 39 | +<div style="background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 8px; padding: 20px; margin: 16px 0;"> |
| 40 | + |
| 41 | +### 📊 Performance Benchmark Comparison |
| 42 | + |
| 43 | +<div style="overflow-x: auto; margin: 16px 0;"> |
| 44 | +<table style="width: 100%; border-collapse: collapse; background: white; border-radius: 6px; overflow: hidden; box-shadow: 0 1px 3px rgba(0,0,0,0.1);"> |
| 45 | + <thead style="background: #f8fafc;"> |
| 46 | + <tr> |
| 47 | + <th style="padding: 12px; text-align: left; border-bottom: 1px solid #e2e8f0;">Model</th> |
| 48 | + <th style="padding: 12px; text-align: left; border-bottom: 1px solid #e2e8f0;">Benchmark Model</th> |
| 49 | + <th style="padding: 12px; text-align: left; border-bottom: 1px solid #e2e8f0;">Hardware Requirements</th> |
| 50 | + <th style="padding: 12px; text-align: left; border-bottom: 1px solid #e2e8f0;">License</th> |
| 51 | + </tr> |
| 52 | + </thead> |
| 53 | + <tbody> |
| 54 | + <tr> |
| 55 | + <td style="padding: 12px; border-bottom: 1px solid #f1f5f9;"><strong>GPT-OSS-120B</strong></td> |
| 56 | + <td style="padding: 12px; border-bottom: 1px solid #f1f5f9;">OpenAI o4-mini</td> |
| 57 | + <td style="padding: 12px; border-bottom: 1px solid #f1f5f9;">Single 80GB GPU</td> |
| 58 | + <td style="padding: 12px; border-bottom: 1px solid #f1f5f9;">Apache 2.0</td> |
| 59 | + </tr> |
| 60 | + <tr> |
| 61 | + <td style="padding: 12px; border-bottom: 1px solid #f1f5f9;"><strong>GPT-OSS-20B</strong></td> |
| 62 | + <td style="padding: 12px; border-bottom: 1px solid #f1f5f9;">OpenAI o3-mini</td> |
| 63 | + <td style="padding: 12px; border-bottom: 1px solid #f1f5f9;">16GB Memory Device</td> |
| 64 | + <td style="padding: 12px; border-bottom: 1px solid #f1f5f9;">Apache 2.0</td> |
| 65 | + </tr> |
| 66 | + </tbody> |
| 67 | +</table> |
| 68 | +</div> |
| 69 | + |
| 70 | +</div> |
| 71 | + |
| 72 | +## 🎯 Application Scenarios |
| 73 | + |
| 74 | +<div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(250px, 1fr)); gap: 16px; margin: 16px 0;"> |
| 75 | + |
| 76 | +<div style="background: #dcfce7; border-left: 4px solid #059669; padding: 16px; border-radius: 4px;"> |
| 77 | +<strong>📱 On-Device Applications</strong><br> |
| 78 | +<ul style="margin: 8px 0 0 0; padding-left: 20px; color: #065f46;"> |
| 79 | + <li>Mobile AI assistants</li> |
| 80 | + <li>Offline intelligent applications</li> |
| 81 | + <li>Edge computing scenarios</li> |
| 82 | +</ul> |
| 83 | +</div> |
| 84 | + |
| 85 | +<div style="background: #fef3c7; border-left: 4px solid #f59e0b; padding: 16px; border-radius: 4px;"> |
| 86 | +<strong>🔧 Development & Research</strong><br> |
| 87 | +<ul style="margin: 8px 0 0 0; padding-left: 20px; color: #92400e;"> |
| 88 | + <li>Rapid prototyping</li> |
| 89 | + <li>Local inference testing</li> |
| 90 | + <li>Low-cost experimentation</li> |
| 91 | +</ul> |
| 92 | +</div> |
| 93 | + |
| 94 | +<div style="background: #eff6ff; border-left: 4px solid #2563eb; padding: 16px; border-radius: 4px;"> |
| 95 | +<strong>🏢 Enterprise Deployment</strong><br> |
| 96 | +<ul style="margin: 8px 0 0 0; padding-left: 20px; color: #1e40af;"> |
| 97 | + <li>Private deployment</li> |
| 98 | + <li>Data security protection</li> |
| 99 | + <li>Cost optimization</li> |
| 100 | +</ul> |
| 101 | +</div> |
| 102 | + |
| 103 | +</div> |
| 104 | + |
| 105 | +## 🏅 Benchmark Performance |
| 106 | + |
| 107 | +<div style="background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 8px; padding: 20px; margin: 16px 0;"> |
| 108 | + |
| 109 | +### 🎯 Core Capability Assessment |
| 110 | + |
| 111 | +<div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(200px, 1fr)); gap: 12px; margin: 16px 0;"> |
| 112 | + <div style="background: white; padding: 12px; border-radius: 6px; border: 1px solid #e2e8f0; text-align: center;"> |
| 113 | + <span style="background: #dcfce7; color: #059669; padding: 4px 12px; border-radius: 12px; font-size: 12px;">✅ Tool Usage</span> |
| 114 | + </div> |
| 115 | + <div style="background: white; padding: 12px; border-radius: 6px; border: 1px solid #e2e8f0; text-align: center;"> |
| 116 | + <span style="background: #dcfce7; color: #059669; padding: 4px 12px; border-radius: 12px; font-size: 12px;">✅ Few-shot Function Calling</span> |
| 117 | + </div> |
| 118 | + <div style="background: white; padding: 12px; border-radius: 6px; border: 1px solid #e2e8f0; text-align: center;"> |
| 119 | + <span style="background: #dcfce7; color: #059669; padding: 4px 12px; border-radius: 12px; font-size: 12px;">✅ CoT Reasoning</span> |
| 120 | + </div> |
| 121 | + <div style="background: white; padding: 12px; border-radius: 6px; border: 1px solid #e2e8f0; text-align: center;"> |
| 122 | + <span style="background: #dcfce7; color: #059669; padding: 4px 12px; border-radius: 12px; font-size: 12px;">✅ Agent Evaluation</span> |
| 123 | + </div> |
| 124 | +</div> |
| 125 | + |
| 126 | +<div style="background: #dcfce7; border-left: 4px solid #059669; padding: 16px; margin: 16px 0; border-radius: 4px;"> |
| 127 | +<strong>🏆 Outstanding Performance</strong><br> |
| 128 | +Strong performance in Tau-Bench agent evaluation suite and HealthBench testing, even surpassing proprietary models like OpenAI o1 and GPT-4o. |
| 129 | +</div> |
| 130 | + |
| 131 | +</div> |
| 132 | + |
| 133 | +## 📋 Technical Specifications |
| 134 | + |
| 135 | +<div style="background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 8px; padding: 20px; margin: 16px 0;"> |
| 136 | + |
| 137 | +<div style="display: grid; grid-template-columns: 1fr 1fr; gap: 20px;"> |
| 138 | + |
| 139 | +<div> |
| 140 | +<h4 style="margin-top: 0; color: #1e40af;">🔧 GPT-OSS-120B</h4> |
| 141 | +<ul style="margin: 0; padding-left: 20px;"> |
| 142 | + <li><strong>Parameter Scale</strong>: 120 billion parameters</li> |
| 143 | + <li><strong>Recommended Hardware</strong>: 80GB GPU</li> |
| 144 | + <li><strong>Performance Benchmark</strong>: OpenAI o4-mini</li> |
| 145 | + <li><strong>Use Cases</strong>: High-performance reasoning tasks</li> |
| 146 | +</ul> |
| 147 | +</div> |
| 148 | + |
| 149 | +<div> |
| 150 | +<h4 style="margin-top: 0; color: #1e40af;">⚡ GPT-OSS-20B</h4> |
| 151 | +<ul style="margin: 0; padding-left: 20px;"> |
| 152 | + <li><strong>Parameter Scale</strong>: 20 billion parameters</li> |
| 153 | + <li><strong>Minimum Configuration</strong>: 16GB memory</li> |
| 154 | + <li><strong>Performance Benchmark</strong>: OpenAI o3-mini</li> |
| 155 | + <li><strong>Use Cases</strong>: Edge device deployment</li> |
| 156 | +</ul> |
| 157 | +</div> |
| 158 | + |
| 159 | +</div> |
| 160 | + |
| 161 | +</div> |
| 162 | + |
| 163 | +## 🚀 Getting Started |
| 164 | + |
| 165 | +<div style="background: linear-gradient(135deg, #eff6ff, #dbeafe); border: 1px solid #bfdbfe; border-radius: 8px; padding: 20px; margin: 16px 0;"> |
| 166 | + |
| 167 | +<h3 style="margin-top: 0; color: #1e40af;">📦 Quick Deployment</h3> |
| 168 | + |
| 169 | +<div style="background: #eff6ff; border-left: 4px solid #2563eb; padding: 16px; margin: 16px 0; border-radius: 4px;"> |
| 170 | +<strong>📄 License Advantages</strong><br> |
| 171 | +Apache 2.0 license provides maximum usage flexibility, supporting commercial applications, modifications, and distribution without license restrictions. |
| 172 | +</div> |
| 173 | + |
| 174 | +<div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(200px, 1fr)); gap: 12px; margin: 16px 0;"> |
| 175 | + <div style="background: white; padding: 12px; border-radius: 6px; border: 1px solid #e2e8f0; text-align: center;"> |
| 176 | + <span style="background: #dbeafe; color: #1e40af; padding: 4px 12px; border-radius: 12px; font-size: 12px;">🔓 Open Source & Free</span> |
| 177 | + </div> |
| 178 | + <div style="background: white; padding: 12px; border-radius: 6px; border: 1px solid #e2e8f0; text-align: center;"> |
| 179 | + <span style="background: #dbeafe; color: #1e40af; padding: 4px 12px; border-radius: 12px; font-size: 12px;">💼 Commercial Friendly</span> |
| 180 | + </div> |
| 181 | + <div style="background: white; padding: 12px; border-radius: 6px; border: 1px solid #e2e8f0; text-align: center;"> |
| 182 | + <span style="background: #dbeafe; color: #1e40af; padding: 4px 12px; border-radius: 12px; font-size: 12px;">🔧 Customizable</span> |
| 183 | + </div> |
| 184 | + <div style="background: white; padding: 12px; border-radius: 6px; border: 1px solid #e2e8f0; text-align: center;"> |
| 185 | + <span style="background: #dbeafe; color: #1e40af; padding: 4px 12px; border-radius: 12px; font-size: 12px;">🌐 Community Support</span> |
| 186 | + </div> |
| 187 | +</div> |
| 188 | + |
| 189 | +</div> |
| 190 | + |
| 191 | +## 📖 User Guide |
| 192 | + |
| 193 | +<div style="background: #eff6ff; border-left: 4px solid #2563eb; padding: 16px; margin: 16px 0; border-radius: 4px;"> |
| 194 | + <strong>💡 Quick Start</strong><br> |
| 195 | + After completing the model deployment, you can view the model usage instructions on the Computing Nest service instance overview page, which provides API call examples, internal network access addresses, public network access addresses, and ApiKey. |
| 196 | +</div> |
| 197 | + |
| 198 | +<div style="background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 8px; padding: 20px; margin: 16px 0;"> |
| 199 | + <div style="text-align: center; margin-bottom: 16px;"> |
| 200 | + <img src="../image-en/img-llm-use-desc.png" alt="Model usage instructions interface" style="max-width: 100%; border-radius: 6px; box-shadow: 0 2px 8px rgba(0,0,0,0.1);"> |
| 201 | + </div> |
| 202 | +</div> |
| 203 | + |
| 204 | +### 🔌 API Call Methods |
| 205 | + |
| 206 | +#### 🖥️ Curl Command Call |
| 207 | + |
| 208 | +<div style="background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 8px; padding: 20px; margin: 16px 0;"> |
| 209 | + |
| 210 | +<div style="text-align: center; margin-bottom: 16px;"> |
| 211 | + <img src="../image-en/img-api-call.png" alt="API call example" style="max-width: 100%; border-radius: 6px; box-shadow: 0 2px 8px rgba(0,0,0,0.1);"> |
| 212 | +</div> |
| 213 | + |
| 214 | +<div style="background: #eff6ff; border-left: 4px solid #2563eb; padding: 16px; margin: 16px 0; border-radius: 4px;"> |
| 215 | + <strong>📋 Parameter Description</strong><br> |
| 216 | + • <code>${ServerIP}</code>: IP address from internal or public network address<br> |
| 217 | + • <code>${ApiKey}</code>: ApiKey provided on the page<br> |
| 218 | + • <code>${ModelName}</code>: Model name |
| 219 | +</div> |
| 220 | + |
| 221 | +Curl command calls can directly use the API call examples from the service instance overview page. The specific structure for calling the model API is as follows: |
| 222 | + |
| 223 | +```bash |
| 224 | +curl -X Post http://${ServerIP}:8000/v1/chat/completions \ |
| 225 | + -H "Content-Type: application/json" \ |
| 226 | + -H "Authorization: Bearer ${ApiKey}" \ |
| 227 | + -d '{ |
| 228 | + "model": "${ModelName}", |
| 229 | + "messages": [ |
| 230 | + { |
| 231 | + "role": "user", |
| 232 | + "content": "Write a letter to my daughter from the future year 2035, telling her to study technology well, become the master of technology, and promote technological and economic development; she is currently in 3rd grade" |
| 233 | + } |
| 234 | + ] |
| 235 | + }' |
| 236 | +``` |
| 237 | + |
| 238 | +</div> |
| 239 | + |
| 240 | +#### 🐍 Python SDK Call |
| 241 | + |
| 242 | +<div style="background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 8px; padding: 20px; margin: 16px 0;"> |
| 243 | + |
| 244 | +<div style="background: #eff6ff; border-left: 4px solid #2563eb; padding: 16px; margin: 16px 0; border-radius: 4px;"> |
| 245 | + <strong>⚙️ Configuration Instructions</strong><br> |
| 246 | + • <code>${ApiKey}</code>: Fill in the ApiKey from the page<br> |
| 247 | + • <code>${ServerUrl}</code>: Fill in the public or internal network address from the page, must include <code>/v1</code> |
| 248 | +</div> |
| 249 | + |
| 250 | +The following is Python example code: |
| 251 | + |
| 252 | +```python |
| 253 | +from openai import OpenAI |
| 254 | + |
| 255 | +##### API Configuration ##### |
| 256 | +openai_api_key = "${ApiKey}" |
| 257 | +openai_api_base = "${ServerUrl}" |
| 258 | + |
| 259 | +client = OpenAI( |
| 260 | + api_key=openai_api_key, |
| 261 | + base_url=openai_api_base, |
| 262 | +) |
| 263 | + |
| 264 | +models = client.models.list() |
| 265 | +model = models.data[0].id |
| 266 | +print(model) |
| 267 | + |
| 268 | + |
| 269 | +def main(): |
| 270 | + stream = True |
| 271 | + |
| 272 | + chat_completion = client.chat.completions.create( |
| 273 | + messages=[ |
| 274 | + { |
| 275 | + "role": "user", |
| 276 | + "content": [ |
| 277 | + { |
| 278 | + "type": "text", |
| 279 | + "text": "Hello, please introduce yourself in as much detail as possible.", |
| 280 | + } |
| 281 | + ], |
| 282 | + } |
| 283 | + ], |
| 284 | + model=model, |
| 285 | + max_completion_tokens=1024, |
| 286 | + stream=stream, |
| 287 | + ) |
| 288 | + |
| 289 | + if stream: |
| 290 | + for chunk in chat_completion: |
| 291 | + print(chunk.choices[0].delta.content, end="") |
| 292 | + else: |
| 293 | + result = chat_completion.choices[0].message.content |
| 294 | + print(result) |
| 295 | + |
| 296 | + |
| 297 | +if __name__ == "__main__": |
| 298 | + main() |
| 299 | +``` |
| 300 | + |
| 301 | +</div> |
| 302 | + |
| 303 | +### 🌐 Web Application Access |
| 304 | + |
| 305 | +<div style="background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 8px; padding: 20px; margin: 16px 0;"> |
| 306 | + |
| 307 | +#### 📱 Access Steps |
| 308 | + |
| 309 | +<div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(250px, 1fr)); gap: 16px; margin: 16px 0;"> |
| 310 | + |
| 311 | +<div style="background: #eff6ff; border-left: 4px solid #2563eb; padding: 16px; border-radius: 4px;"> |
| 312 | +<h4 style="margin-top: 0; color: #1e40af;">🔗 Step 1: Get Access Link</h4> |
| 313 | +<p style="margin: 0;">On the service instance overview page, click the link corresponding to the Web application to directly access the model service Web interface.</p> |
| 314 | +</div> |
| 315 | + |
| 316 | +<div style="background: #dcfce7; border-left: 4px solid #059669; padding: 16px; border-radius: 4px;"> |
| 317 | +<h4 style="margin-top: 0; color: #059669;">💬 Step 2: Start Conversation</h4> |
| 318 | +<p style="margin: 0;">Enter your question in the input box on the model service Web page to start conversing with the large language model.</p> |
| 319 | +</div> |
| 320 | + |
| 321 | +</div> |
| 322 | + |
| 323 | +#### 🖼️ Interface Display |
| 324 | + |
| 325 | +<div style="text-align: center; margin: 20px 0;"> |
| 326 | + <img src="../image-en/img-web.png" alt="Web Application Access Entry" style="max-width: 100%; border-radius: 8px; box-shadow: 0 4px 8px rgba(0,0,0,0.1); border: 1px solid #e2e8f0;"> |
| 327 | +</div> |
| 328 | + |
| 329 | +<div style="background: #eff6ff; border-left: 4px solid #2563eb; padding: 16px; margin: 16px 0; border-radius: 4px;"> |
| 330 | + <strong>💡 Access Tips</strong><br> |
| 331 | + Find the link corresponding to the Web application on the service instance overview page and click it to directly access the model service Web interface. |
| 332 | +</div> |
| 333 | + |
| 334 | +<div style="text-align: center; margin: 20px 0;"> |
| 335 | + <img src="../image-en/img-appflow.png" alt="Model Conversation Interface" style="max-width: 100%; border-radius: 8px; box-shadow: 0 4px 8px rgba(0,0,0,0.1); border: 1px solid #e2e8f0;"> |
| 336 | +</div> |
| 337 | + |
| 338 | +<div style="background: #dcfce7; border-left: 4px solid #059669; padding: 16px; margin: 16px 0; border-radius: 4px;"> |
| 339 | + <strong>✅ Usage Instructions</strong><br> |
| 340 | + Enter your questions or requirements in the input box, and the system will respond in real-time and provide corresponding model services. |
| 341 | +</div> |
| 342 | + |
| 343 | +</div> |
| 344 | + |
| 345 | +--- |
| 346 | + |
| 347 | +<div style="text-align: center; padding: 16px; background: #f8fafc; border-radius: 6px; margin-top: 24px;"> |
| 348 | + <p style="margin: 0; color: #64748b; font-size: 14px;"> |
| 349 | + 🚀 <strong>GPT-OSS Series</strong> | The Future of Open AI, Powerful Performance Within Reach |
| 350 | + </p> |
| 351 | +</div> |
0 commit comments