[Pytorch][AO] Update choose_qparams_per_token op to output correct shape for scales and zp #136807
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
…ape for scales and zp - also makes scales and zp dtype reconcile with meta impl as well as other quantized ops representation of scales and zero point - make sure qunatize_per_token's output_dtype is respected There are a few places where we need to reconcile on scale and zero point dtype but that will come later. This fixes are mainly being done to enable quantized kv cache though ET stack Differential Revision: [D62301840](https://our.internmc.facebook.com/intern/diff/D62301840/) [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/136807
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 Cancelled JobAs of commit 4e93552 with merge base 2421344 ( CANCELLED JOB - The following job was cancelled. Please retry:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
pytorch-bot
bot
added
release notes: quantization
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D62301840 |
jerryzh168
approved these changes
Sep 27, 2024
jerryzh168
reviewed
Sep 27, 2024
| input (torch.Tensor): quantized Tensor (uint8, int8 etc.) | ||
| scales (float32 torch.Tensor): quantization parameter for per token affine quantization | ||
| scales (float64 torch.Tensor): quantization parameter for per token affine quantization | ||
| zero_points (int32 torch.Tensor): quantization parameter for per token affine quantization |
Contributor
There was a problem hiding this comment.
should zero_points be int64?
… correct shape for scales and zp" - also makes scales and zp dtype reconcile with meta impl as well as other quantized ops representation of scales and zero point - make sure qunatize_per_token's output_dtype is respected There are a few places where we need to reconcile on scale and zero point dtype but that will come later. This fixes are mainly being done to enable quantized kv cache though ET stack Differential Revision: [D62301840](https://our.internmc.facebook.com/intern/diff/D62301840/) [ghstack-poisoned]
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D62301840 |
kimishpatel
added a commit
that referenced
this pull request
Sep 27, 2024
…ape for scales and zp Pull Request resolved: #136807 - also makes scales and zp dtype reconcile with meta impl as well as other quantized ops representation of scales and zero point - make sure qunatize_per_token's output_dtype is respected There are a few places where we need to reconcile on scale and zero point dtype but that will come later. This fixes are mainly being done to enable quantized kv cache though ET stack Differential Revision: [D62301840](https://our.internmc.facebook.com/intern/diff/D62301840/) ghstack-source-id: 245083869
Contributor
|
@pytorchbot merge -f 'Landed internally' (Initiating merge automatically since Phabricator Diff has merged, using force because this PR might not pass merge_rules.json but landed internally) |
Collaborator
Merge startedYour change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Please use Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
jerryzh168
added a commit
to jerryzh168/ao
that referenced
this pull request
Oct 3, 2024
Summary: Fixes: pytorch#970 The test was broken by a recent refactor in pytorch: pytorch/pytorch#136807 Test Plan: CI Reviewers: Subscribers: Tasks: Tags:
jerryzh168
added a commit
to jerryzh168/ao
that referenced
this pull request
Oct 4, 2024
Summary: Fixes: pytorch#970 The test was broken by a recent refactor in pytorch: pytorch/pytorch#136807 Test Plan: CI Reviewers: Subscribers: Tasks: Tags:
msaroufim
pushed a commit
to pytorch/ao
that referenced
this pull request
Oct 4, 2024
* Unskip `test_choose_qparams_token_asym` in 2.6 Summary: Fixes: #970 The test was broken by a recent refactor in pytorch: pytorch/pytorch#136807 Test Plan: CI Reviewers: Subscribers: Tasks: Tags: * fix
melvinebenezer
pushed a commit
to melvinebenezer/ao
that referenced
this pull request
Oct 7, 2024
* Unskip `test_choose_qparams_token_asym` in 2.6 Summary: Fixes: pytorch#970 The test was broken by a recent refactor in pytorch: pytorch/pytorch#136807 Test Plan: CI Reviewers: Subscribers: Tasks: Tags: * fix
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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.
Stack from ghstack (oldest at bottom):
quantized ops representation of scales and zero point
There are a few places where we need to reconcile on scale and zero point dtype
but that will come later. This fixes are mainly being done to enable quantized
kv cache though ET stack
Differential Revision: D62301840