[bc-breaking] Generalize QAT configs beyond intx quantization #2608
Conversation
This file contains hidden or 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
**Summary:** Current QAT APIs are highly tailored to intx quantization, but we plan to extend QAT to other dtypes such as fp8 and nvfp4 in the near future. This is the necessary refactor before that. In particular, we are moving references to integer quantization from the main `IntXQuantizationAwareTrainingConfig` to the inner `FakeQuantizeConfig`s: ``` IntXQuantizationAwareTrainingConfig -> QuantizationAwareTrainingConfig FakeQuantizeConfig -> IntxFakeQuantizeConfig ``` In the future, we will have other types of FakeQuantizeConfig for float dtypes that users can pass in instead of the existing Intx one. **BC-breaking notes:** For BC, we keep around the old names to reference the new ones. However, this commit is still BC-breaking in the sense that a few APIs now accept the abstract `FakeQuantizeConfigBase` instead. For the most part, this abstract class will be hidden from the user. Before: ``` activation_config = FakeQuantizeConfig(torch.int8, "per_token", is_symmetric=False) weight_config = FakeQuantizeConfig(torch.int4, group_size=32) qat_config = IntXQuantizationAwareTrainingConfig(activation_config, weight_config), quantize_(model, qat_config) train(model) quantize_(model, FromIntXQuantizationAwareTrainingConfig()) quantize_(model, Int8DynamicActivationInt4WeightConfig(group_size=32)) ``` After: ``` activation_config = IntxFakeQuantizeConfig(torch.int8, "per_token", is_symmetric=False) weight_config = IntxFakeQuantizeConfig(torch.int4, group_size=32) qat_config = QuantizationAwareTrainingConfig(activation_config, weight_config), quantize_(model, qat_config) train(model) quantize_(model, FromQuantizationAwareTrainingConfig()) quantize_(model, Int8DynamicActivationInt4WeightConfig(group_size=32)) ``` **Test Plan:** python test/quantization/test_qat.py
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/2608
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 2125f2f with merge base 12ff479 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
andrewor14
added
the
topic: bc-breaking
|
since we are planning to break BC, I'm wondering if is it possible to merge or just to extend |
|
Changing the design, closing this for now. Thanks @jerryzh168 |
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.
Summary: Current QAT APIs are highly tailored to intx quantization, but we plan to extend QAT to other dtypes such as fp8 and nvfp4 in the near future. This is the necessary refactor before that.
In particular, we are moving references to integer quantization from the main
IntXQuantizationAwareTrainingConfigto the innerFakeQuantizeConfigs:In the future, we will have other types of FakeQuantizeConfig for float dtypes that users can pass in instead of the existing Intx one.
BC-breaking notes: For BC, we keep around the old names to reference the new ones. However, this commit is still BC-breaking in the sense that a few APIs now accept the abstract
FakeQuantizeConfigBaseinstead. For the most part, this abstract class will be hidden from the user.Before:
After:
Test Plan:
python test/quantization/test_qat.py