-
Notifications
You must be signed in to change notification settings - Fork 25.3k
[quant][pt2e] Add reference representation for quantized conv2d #105784
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/105784
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: ✅ No FailuresAs of commit 6b3623b: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 020000c Pull Request resolved: #105784
…onv2d" Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 94e3440 Pull Request resolved: #105784
…onv2d" Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: b4b33d8 Pull Request resolved: #105784
…onv2d" Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
…onv2d" Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
…onv2d" Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
…onv2d" Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: c953a3f Pull Request resolved: #105784
…onv2d" Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
…onv2d" Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 938366d Pull Request resolved: #105784
…onv2d" Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 2c98ccf Pull Request resolved: #105784
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
unblock:
please add comments, that we landing this to unblock but this needs to be vetted since these are not functionally tested and some comments are not addressed
I have tested these locally and fixed the bias stuff as well, will probably enable the tests in CI soon since I didn't find it is timing out in sandcastle recently. what is the not addressed comment are you referring to btw? |
…onv2d" Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 04fabef Pull Request resolved: #105784
@pytorchbot merge |
TODO: we need to support autograd for out_dtype op so that user don't need to use torch.no_grad to run the quantized model. |
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 |
…rch#105784) Summary: Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8 Test Plan: python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel Although right now it is not really testing things since there is some problem with dynamo export Summary: Test Plan: Reviewers: Subscribers: Tasks: Tags: Pull Request resolved: pytorch#105784 Approved by: https://github.com/kimishpatel ghstack dependencies: pytorch#105783
Stack from ghstack (oldest at bottom):
Summary:
Implementing reference representation for quantized ops we decided in https://docs.google.com/document/d/17h-OEtD4o_hoVuPqUFsdm5uo7psiNMY8ThN03F9ZZwg/edit#heading=h.ov8z39149wy8
Test Plan:
python test/test_quantization.py TestQuantizePT2E.test_representation_quantize_dequantize_per_channel
Although right now it is not really testing things since there is some problem with dynamo export
Summary:
Test Plan:
Reviewers:
Subscribers:
Tasks:
Tags: