Add new INT4 quantization features to model builder #940
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
yufenglee
reviewed
Sep 30, 2024
| "accuracy_level": int(extra_options["int4_accuracy_level"]) if "int4_accuracy_level" in extra_options else 0, # Default is 0 for non-QDQ formats, default is 4 for QDQ formats | ||
| "block_size": int(extra_options["int4_block_size"]) if "int4_block_size" in extra_options else 32, | ||
| "op_types_to_quantize": (extra_options["op_type_to_quantize"], ) if "int4_op_type_to_quantize" in extra_options else ("MatMul", "Gather", ), |
There was a problem hiding this comment.
Could we make the default "MatMul"? The reason is that the quantization of Gather is optional and not all the EPs support it.
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Outdated suggestions cannot be applied.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.
Description
This PR adds new INT4 quantization features to the model builder.
MatMulcan now be quantized toMatMulNBitsvia RTN.int4_op_type_to_quantizehas been added to allow more flexibility with INT4 quantization.Motivation and Context
With these PR changes, the size of the ONNX models can be reduced by quantizing the embedding layer and/or the language modeling head.
For the ONNX models built from already-quantized PyTorch models, one example is with using AutoAWQ. AutoAWQ does not quantize the language modeling head. The resulting ONNX model typically contains a
MatMulop for the language modeling head. Now, thatMatMulop will be quantized via RTN toMatMulNBitsto reduce memory.