Load suno/bark failed due to safetensor #42541
Description
System Info
latest version of transformers.
Who can help?
Information
- The official example scripts
- My own modified scripts
Tasks
- An officially supported task in the
examplesfolder (such as GLUE/SQuAD, ...) - My own task or dataset (give details below)
Reproduction
from transformers import pipeline
synthesiser = pipeline("text-to-speech", "suno/bark")
output:
Error during conversion: ReadTimeout('The read operation timed out')
Error during conversion: AttributeError("'str' object has no attribute 'decode'")
Error during conversion: AttributeError("'str' object has no attribute 'decode'")
Traceback (most recent call last):
File "/home/jiqing/transformers/src/transformers/pipelines/base.py", line 248, in load_model
model = model_class.from_pretrained(model, **fp32_kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/jiqing/transformers/src/transformers/modeling_utils.py", line 250, in _wrapper
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "/home/jiqing/transformers/src/transformers/modeling_utils.py", line 3937, in from_pretrained
checkpoint_files, sharded_metadata = _get_resolved_checkpoint_files(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/jiqing/transformers/src/transformers/modeling_utils.py", line 547, in _get_resolved_checkpoint_files
resolved_archive_file, revision, is_sharded = auto_conversion(
^^^^^^^^^^^^^^^^
File "/home/jiqing/transformers/src/transformers/safetensors_conversion.py", line 110, in auto_conversion
raise e
File "/home/jiqing/transformers/src/transformers/safetensors_conversion.py", line 89, in auto_conversion
sha = get_conversion_pr_reference(api, pretrained_model_name_or_path, **cached_file_kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/jiqing/transformers/src/transformers/safetensors_conversion.py", line 76, in get_conversion_pr_reference
raise OSError(
OSError: Could not create safetensors conversion PR. The repo does not appear to have a file named pytorch_model.bin or model.safetensors.If you are loading with variant, use `use_safetensors=False` to load the original model.
Expected behavior
Regression PR #41580
I submit the suno/bark model in https://safetensors-convert.hf.space/ , but got
Model suno/bark already has an open PR check out https://huggingface.co/suno/bark/discussions/37