search

facebookresearch / esm

Evolutionary Scale Modeling (esm): Pretrained language models for proteins
MIT License
2.97k stars 586 forks source link

HuggingFace EsmForProteinFolding inference raises exception on Windows #647

Open shimooper opened 6 months ago

shimooper commented 6 months ago

Bug description When trying to run inference using EsmForProteinFolding from HuggingFace transformers library on Windows, an exception is raised

Reproduction steps Code: from transformers import AutoTokenizer, EsmForProteinFolding

tokenizer = AutoTokenizer.from_pretrained("facebook/esmfold_v1") model = EsmForProteinFolding.from_pretrained("facebook/esmfold_v1")

inputs = tokenizer("MLKNVQVQLV", return_tensors="pt", add_special_tokens=False) # A tiny random peptide

outputs = model(**inputs)

Logs Traceback (most recent call last): File "C:\Users\yairs\OneDrive\Documents\University\Master\stuff\python packages tutorials\esm\EsmForProteinFolding.py", line 8, in outputs = model(inputs) File "C:\Users\yairs\miniconda3\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl return self._call_impl(*args, *kwargs) File "C:\Users\yairs\miniconda3\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl return forward_call(args, kwargs) File "C:\Users\yairs\miniconda3\lib\site-packages\transformers\models\esm\modeling_esmfold.py", line 2156, in forward structure: dict = self.trunk(s_s_0, s_z_0, aa, position_ids, attention_mask, no_recycles=num_recycles) File "C:\Users\yairs\miniconda3\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl return self._call_impl(*args, kwargs) File "C:\Users\yairs\miniconda3\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl return forward_call(*args, *kwargs) File "C:\Users\yairs\miniconda3\lib\site-packages\transformers\models\esm\modeling_esmfold.py", line 1965, in forward structure = self.structure_module( File "C:\Users\yairs\miniconda3\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl return self._call_impl(args, kwargs) File "C:\Users\yairs\miniconda3\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl return forward_call(*args, **kwargs) File "C:\Users\yairs\miniconda3\lib\site-packages\transformers\models\esm\modeling_esmfold.py", line 1799, in forward pred_xyz = self.frames_and_literature_positions_to_atom14_pos(all_frames_to_global, aatype) File "C:\Users\yairs\miniconda3\lib\site-packages\transformers\models\esm\modeling_esmfold.py", line 1880, in frames_and_literature_positions_to_atom14_pos return frames_and_literature_positions_to_atom14_pos( File "C:\Users\yairs\miniconda3\lib\site-packages\transformers\models\esm\openfold_utils\feats.py", line 236, in frames_and_literature_positions_to_atom14_pos group_mask_one_hot: torch.LongTensor = nn.functional.one_hot( RuntimeError: one_hot is only applicable to index tensor.

Huangq0825 commented 1 month ago

the same error!!! did you solve it?