Update _base.py - Remote (.bin) model load fix #465
Merged
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Whats happening -
.bin
Model is unable to load - FIleNotFound error.When you load a model from huggingface remotely with
use_safetensors = False
, there are two extensions to work with -.bin
,.pt
.Line 834, checks the cached model directory for the model base name with the above extensions in the same order as above.
Case - 1: If no model found with either extensions, the original code works as expected (throws error).
Case - 2: If models with both extensions found or if only
.pt
extension is found, the model with.pt
extension loads. Although in the former case, the user has no liberty to choose between the two models.Case - 3: If only
.bin
model exists. then it throws a FileNotFoundError as in the first case.The fix -
A simple fix with
temp
variable, which hold the latest found model in order to prevent theresolved_archive_file
variable defaulting toNone
incase the model with next search extension is not found.