Fix GPTQ for models which do not have float16 at the default dtype (simpler) #1953
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
5 tasks
Before this change GPTQ models would not work if the model's default data type is not `float16`. For example, Gemma GPTQ models would fail because the default dtype of Gemma is `bfloat16`. There are two issues: If the default `dtype` is not `float16`, the quantizer's `float16` parameters get converted to that dtype. The kernels cannot deal with non-`float16` types. The same applies to inputs of quantized ops. This is resolved by setting the dtype of gptq/awq-quantized models to `float16`.
danieldk
force-pushed
the
bugfix/preserve-quantized-dtype-simpler
branch
from
9ce8640 to
6f30a13
Compare
Narsil
approved these changes
May 27, 2024
yuanwu2017
pushed a commit
to yuanwu2017/tgi-gaudi
that referenced
this pull request
Sep 26, 2024
…impler) (huggingface#1953) # What does this PR do? Fix GPTQ for models which do not have float16 at the default dtype Before this change GPTQ models would not work if the model's default data type is not `float16`. For example, Gemma GPTQ models would fail because the default dtype of Gemma is `bfloat16`. There are two issues: If the default `dtype` is not `float16`, the quantizer's `float16` parameters get converted to that dtype. The kernels cannot deal with non-`float16` types. The same applies to inputs of quantized ops. This is resolved by setting the dtype of gptq/awq-quantized models to `float16`. Simpler version of huggingface#1951. **Draft:** just testing... ## Before submitting - [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case). - [x] Did you read the [contributor guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests), Pull Request section? - [ ] Was this discussed/approved via a Github issue or the [forum](https://discuss.huggingface.co/)? Please add a link to it if that's the case. - [ ] Did you make sure to update the documentation with your changes? Here are the [documentation guidelines](https://github.com/huggingface/transformers/tree/main/docs), and [here are tips on formatting docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation). - [ ] Did you write any new necessary tests? ## Who can review? Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR. <!-- Your PR will be replied to more quickly if you can figure out the right person to tag with @ @OlivierDehaene OR @Narsil -->
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.
What does this PR do?
Fix GPTQ for models which do not have float16 at the default dtype
Before this change GPTQ models would not work if the model's default
data type is not
float16. For example, Gemma GPTQ models would failbecause the default dtype of Gemma is
bfloat16. There are two issues:If the default
dtypeis notfloat16, the quantizer'sfloat16parameters get converted to that dtype. The kernels cannot deal
with non-
float16types. The same applies to inputs of quantized ops.This is resolved by setting the dtype of gptq/awq-quantized models to
float16.Simpler version of #1951.
Draft: just testing...
Before submitting
Pull Request section?
to it if that's the case.
documentation guidelines, and
here are tips on formatting docstrings.
Who can review?
Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.