Skip to content

Better HF integration for MambaLMHeadModel #471

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 8 commits into
base: main
Choose a base branch
from
Open

Better HF integration for MambaLMHeadModel #471

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
404df4d
Support HF integration in Mamba and Mamba2Simple + add metadata
Wauplin Jul 16, 2024
44458c4
Add huggingface_hub as core dependency
Wauplin Jul 16, 2024
961eccb
mamba-ssm as library name
Wauplin Jul 16, 2024
4bd4af9
requires 0.23.5
Wauplin Jul 16, 2024
a157ec5
Mixin in MambaLMHeadModel only
Wauplin Jul 16, 2024
bc402de
last line
Wauplin Jul 16, 2024
0c4686f
remove mixin from mamba2
Wauplin Jul 16, 2024
e536a97
add pipeline_tag
Wauplin Jul 16, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 8 additions & 4 deletions mamba_ssm/modules/mamba2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,8 +5,8 @@
import torch
import torch.nn as nn
import torch.nn.functional as F

from einops import rearrange, repeat
from huggingface_hub import PyTorchModelHubMixin

try:
from causal_conv1d import causal_conv1d_fn, causal_conv1d_update
Expand All @@ -31,10 +31,14 @@
from mamba_ssm.ops.triton.ssd_combined import mamba_chunk_scan_combined
from mamba_ssm.ops.triton.ssd_combined import mamba_split_conv1d_scan_combined

from huggingface_hub import PyTorchModelHubMixin


class Mamba2(nn.Module, PyTorchModelHubMixin):
class Mamba2(
nn.Module,
PyTorchModelHubMixin,
library_name="mamba_ssm",
repo_url="https://github.com/state-spaces/mamba",
tags=["mamba2", "arXiv:2312.00752", "arXiv:2405.21060"],
):
def __init__(
self,
d_model,
Expand Down
10 changes: 8 additions & 2 deletions mamba_ssm/modules/mamba2_simple.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,8 +4,8 @@
import torch
import torch.nn as nn
import torch.nn.functional as F

from einops import rearrange, repeat
from huggingface_hub import PyTorchModelHubMixin

try:
from causal_conv1d import causal_conv1d_fn
Expand All @@ -21,7 +21,13 @@
from mamba_ssm.ops.triton.ssd_combined import mamba_split_conv1d_scan_combined


class Mamba2Simple(nn.Module):
class Mamba2Simple(
nn.Module,
PyTorchModelHubMixin,
library_name="mamba_ssm",
repo_url="https://github.com/state-spaces/mamba",
tags=["mamba2simple", "arXiv:2312.00752", "arXiv:2405.21060"],
):
def __init__(
self,
d_model,
Expand Down
9 changes: 8 additions & 1 deletion mamba_ssm/modules/mamba_simple.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,7 @@
from torch import Tensor

from einops import rearrange, repeat
from huggingface_hub import PyTorchModelHubMixin

from mamba_ssm.ops.selective_scan_interface import selective_scan_fn, mamba_inner_fn

Expand All @@ -28,7 +29,13 @@
RMSNorm, layer_norm_fn, rms_norm_fn = None, None, None


class Mamba(nn.Module):
class Mamba(
nn.Module,
PyTorchModelHubMixin,
library_name="mamba_ssm",
repo_url="https://github.com/state-spaces/mamba",
tags=["mamba", "arXiv:2312.00752"],
):
def __init__(
self,
d_model,
Expand Down
1 change: 1 addition & 0 deletions setup.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -374,6 +374,7 @@ def run(self):
"einops",
"triton",
"transformers",
"huggingface_hub>=0.22",
# "causal_conv1d>=1.4.0",
],
)