Fix onnx export for quantized MPT models #1646
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Empty dict raises an exception when exporting quantized MPT models because when `state_dict = {}`, the HF loading pipeline will try to initialize the model by calling `_init_weights` for each module. However, for MPT models, this function is not defined and will therefore raise an exception.
@neuralmagic/machine-learning on a side note, I am not really sure what this function is supposed to do. According to the code it will return either `None` (for pruned model) or `{}` (for quantized model), but never the actual loaded `state_dict` as mentioned in the docstring. We might want to rethink whether we need to refactor the function and its usage, or at least to update the docstring?
bfineran
approved these changes
Jul 3, 2023
shubhra
approved these changes
Jul 13, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Empty dict raises an exception when exporting quantized MPT models because when
state_dict = {}, the HF loading pipeline will try to initialize the model by calling_init_weightsfor each module. However, for MPT models, this function is not defined and will therefore raise an exception.@bfineran on a side note, I am not really sure what this function is supposed to do. According to the code it will return either
None(for pruned model) or{}(for quantized model), but never the actual loadedstate_dictas mentioned in the docstring. We might want to rethink whether we need to refactor the function and its usage, or at least to update the docstring because at the moment they don't match.