[quant][pt2e] Support int16 quantization #108453
Closed
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: Previously we can only use native pytorch int dtypes that has corresponding quantized dtypes (e.g. quint8, qint8), this PR removes this assumption in observers/fake_quants so that users can use all pytorch native dtypes (except for int64, we can add it later if need) the main addition here is int16. Test Plan: python test/test_quantization.py TestQuantizePT2E Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/108453
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit e5c0134 with merge base 66af4f6 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Summary: Previously we can only use native pytorch int dtypes that has corresponding quantized dtypes (e.g. quint8, qint8), this PR removes this assumption in observers/fake_quants so that users can use all pytorch native dtypes (except for int64, we can add it later if need) the main addition here is int16. Test Plan: python test/test_quantization.py TestQuantizePT2E Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
jerryzh168
added a commit
that referenced
this pull request
Sep 1, 2023
Summary: Previously we can only use native pytorch int dtypes that has corresponding quantized dtypes (e.g. quint8, qint8), this PR removes this assumption in observers/fake_quants so that users can use all pytorch native dtypes (except for int64, we can add it later if need) the main addition here is int16. Test Plan: python test/test_quantization.py TestQuantizePT2E Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 347f756 Pull Request resolved: #108453
kimishpatel
reviewed
Sep 5, 2023
| scale = max_val_pos / (float(quant_max - quant_min) / 2) | ||
| scale = torch.max(scale, self.eps) | ||
| if self.dtype == torch.quint8: | ||
| if self.dtype in [torch.quint8, torch.uint8]: |
There was a problem hiding this comment.
why did this have to be changed?
There was a problem hiding this comment.
this is to make sure the code works for both quint8 and uint8 dtypes
kimishpatel
requested changes
Sep 5, 2023
| def annotate(self, model: torch.fx.GraphModule) -> torch.fx.GraphModule: | ||
| # using int32 to simulate int16 | ||
| int16_qspec = QuantizationSpec( | ||
| dtype=torch.int32, |
There was a problem hiding this comment.
this says dtype=int32 not int16
There was a problem hiding this comment.
oh nevermind, I see the restriction on the values. I thought int16 is supported torch dtype but not uint16? Then why not use that dtype directly
There was a problem hiding this comment.
oh I think I can use torch.int16 here directly, not sure why I used int32 in the beginning
andrewor14
reviewed
Sep 5, 2023
Summary: Previously we can only use native pytorch int dtypes that has corresponding quantized dtypes (e.g. quint8, qint8), this PR removes this assumption in observers/fake_quants so that users can use all pytorch native dtypes (except for int64, we can add it later if need) the main addition here is int16. Test Plan: python test/test_quantization.py TestQuantizePT2E Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
jerryzh168
added a commit
that referenced
this pull request
Sep 5, 2023
Summary: Previously we can only use native pytorch int dtypes that has corresponding quantized dtypes (e.g. quint8, qint8), this PR removes this assumption in observers/fake_quants so that users can use all pytorch native dtypes (except for int64, we can add it later if need) the main addition here is int16. Test Plan: python test/test_quantization.py TestQuantizePT2E Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: aa88d50 Pull Request resolved: #108453
Summary: Previously we can only use native pytorch int dtypes that has corresponding quantized dtypes (e.g. quint8, qint8), this PR removes this assumption in observers/fake_quants so that users can use all pytorch native dtypes (except for int64, we can add it later if need) the main addition here is int16. Test Plan: python test/test_quantization.py TestQuantizePT2E Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
jerryzh168
added a commit
that referenced
this pull request
Sep 5, 2023
Summary: Previously we can only use native pytorch int dtypes that has corresponding quantized dtypes (e.g. quint8, qint8), this PR removes this assumption in observers/fake_quants so that users can use all pytorch native dtypes (except for int64, we can add it later if need) the main addition here is int16. Test Plan: python test/test_quantization.py TestQuantizePT2E Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 9191a2c Pull Request resolved: #108453
yeah that's correct, we are not really using torch dtypes before in observer or fx flows |
kimishpatel
approved these changes
Sep 5, 2023
Summary: Previously we can only use native pytorch int dtypes that has corresponding quantized dtypes (e.g. quint8, qint8), this PR removes this assumption in observers/fake_quants so that users can use all pytorch native dtypes (except for int64, we can add it later if need) the main addition here is int16. Test Plan: python test/test_quantization.py TestQuantizePT2E Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
jerryzh168
added a commit
that referenced
this pull request
Sep 5, 2023
Summary: Previously we can only use native pytorch int dtypes that has corresponding quantized dtypes (e.g. quint8, qint8), this PR removes this assumption in observers/fake_quants so that users can use all pytorch native dtypes (except for int64, we can add it later if need) the main addition here is int16. Test Plan: python test/test_quantization.py TestQuantizePT2E Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 469244a Pull Request resolved: #108453
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
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):
Summary:
Previously we can only use native pytorch int dtypes that has corresponding quantized dtypes (e.g. quint8, qint8), this
PR removes this assumption in observers/fake_quants so that users can use all pytorch native dtypes (except for int64, we can add it later if need)
the main addition here is int16.
Test Plan:
python test/test_quantization.py TestQuantizePT2E
Reviewers:
Subscribers:
Tasks:
Tags: