docs(Float8DynamicActivationFloat8WeightConfig): mark as inference-only (#4376)#4451
Open
Anai-Guo wants to merge 1 commit into
Open
docs(Float8DynamicActivationFloat8WeightConfig): mark as inference-only (#4376)#4451Anai-Guo wants to merge 1 commit into
Anai-Guo wants to merge 1 commit into
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/4451
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Closes #4376.
Adds an explicit
Note: inference-onlyblock toFloat8DynamicActivationFloat8WeightConfig's docstring so users picking up this config for LoRA-on-quantized-base (or any other training workflow) discover up-front that the backward pass will raiseRuntimeError: derivative for aten::_scaled_mm is not implemented, and don't have to debug it from a crash.This is option (1) from the issue — the lowest-risk path. The training-aware variant in option (2) is a separate, larger change.
What changes
torchao/quantization/quant_api.py— one docstring-only block added toFloat8DynamicActivationFloat8WeightConfig. Mirrors the existingNote:block style onFloat8WeightOnlyConfiga few lines above.The note:
derivative for aten::_scaled_mm is not implemented).torch._scaled_mmhas no autograd kernel).filter_fnpattern they almost certainly want next.torchao.float8.convert_to_float8_trainingfor the actual float8 training path.No code changes — purely a docstring addition that flows through to the generated API docs.
🤖 Generated with Claude Code