-
Notifications
You must be signed in to change notification settings - Fork 463
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
Add BF16 in padded FP8 quantize ops #2010
Conversation
This pull request was exported from Phabricator. Differential Revision: D49166595 |
✅ Deploy Preview for pytorch-fbgemm-docs canceled.
|
This pull request was exported from Phabricator. Differential Revision: D49166595 |
Summary: - Add BF16 support in `FloatToPaddedFP8RowwiseQuantized` and `PaddedFP8RowwiseQuantizedToFloat`. - Refactor `src/quantize_ops/quantize_fp8_rowwise.cu` - Move unit test from `hpc` to `fbgemm_gpu` Reviewed By: summerdengfb, qchip Differential Revision: D49166595
41b735b
to
4a0114d
Compare
This pull request was exported from Phabricator. Differential Revision: D49166595 |
Summary: - Add BF16 support in `FloatToPaddedFP8RowwiseQuantized` and `PaddedFP8RowwiseQuantizedToFloat`. - Refactor `src/quantize_ops/quantize_fp8_rowwise.cu` - Move unit test from `hpc` to `fbgemm_gpu` Reviewed By: summerdengfb, qchip Differential Revision: D49166595
4a0114d
to
59d29f2
Compare
This pull request was exported from Phabricator. Differential Revision: D49166595 |
59d29f2
to
54ed262
Compare
Summary: - Add BF16 support in `FloatToPaddedFP8RowwiseQuantized` and `PaddedFP8RowwiseQuantizedToFloat`. - Refactor `src/quantize_ops/quantize_fp8_rowwise.cu` - Move unit test from `hpc` to `fbgemm_gpu` Reviewed By: summerdengfb, qchip Differential Revision: D49166595
This pull request was exported from Phabricator. Differential Revision: D49166595 |
1 similar comment
This pull request was exported from Phabricator. Differential Revision: D49166595 |
Summary: - Add BF16 support in `FloatToPaddedFP8RowwiseQuantized` and `PaddedFP8RowwiseQuantizedToFloat`. - Refactor `src/quantize_ops/quantize_fp8_rowwise.cu` - Move unit test from `hpc` to `fbgemm_gpu` Reviewed By: jianyuh, summerdengfb, qchip Differential Revision: D49166595
54ed262
to
95a116c
Compare
This pull request was exported from Phabricator. Differential Revision: D49166595 |
1 similar comment
This pull request was exported from Phabricator. Differential Revision: D49166595 |
This pull request has been merged in 3950764. |
Summary:
FloatToPaddedFP8RowwiseQuantized
andPaddedFP8RowwiseQuantizedToFloat
.src/quantize_ops/quantize_fp8_rowwise.cu
hpc
tofbgemm_gpu
Differential Revision: D49166595