-
Notifications
You must be signed in to change notification settings - Fork 536
4b embedding quantizer #3081
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
4b embedding quantizer #3081
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/3081
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 74b78e7 with merge base 1f4b631 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This pull request was exported from Phabricator. Differential Revision: D56229021 |
Summary: 4b embedding quantizer Reviewed By: larryliu0820 Differential Revision: D56229021
8699dc9
to
2f18427
Compare
Summary: 4b embedding quantizer Reviewed By: larryliu0820 Differential Revision: D56229021
2f18427
to
b13d167
Compare
This pull request was exported from Phabricator. Differential Revision: D56229021 |
1 similar comment
This pull request was exported from Phabricator. Differential Revision: D56229021 |
self.weight, self.scales, None, 0, 0, indices, dtype=self.dtype | ||
) | ||
else: # 4bit packed | ||
return torch.ops.llama_quantized.embedding_4bit.dtype( |
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.
It seems here it should be quantized_decomposed::embedding_4bit
) | ||
|
||
|
||
class EmbeddingOnlyInt8QuantHandler: | ||
def __init__(self, mod, *, bitwidth: int = 8, group_size: Optional[int] = None): | ||
class EmbeddingOnlyInt8QuantHandler(QuantHandler): |
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.
Should we rename this class? Since it's not int8 only anymore.
Summary: 4b embedding quantizer Reviewed By: larryliu0820 Differential Revision: D56229021
b13d167
to
74b78e7
Compare
This pull request was exported from Phabricator. Differential Revision: D56229021 |
Summary: 4b embedding quantizer Reviewed By: larryliu0820 Differential Revision: D56229021
Summary: 4b embedding quantizer Reviewed By: larryliu0820 Differential Revision: D56229021
Summary: 4b embedding quantizer Reviewed By: larryliu0820 Differential Revision: D56229021
Summary: 4b embedding quantizer Reviewed By: larryliu0820 Differential Revision: D56229021
Summary: 4b embedding quantizer Reviewed By: larryliu0820 Differential Revision: D56229021
Summary: 4b embedding quantizer Reviewed By: larryliu0820 Differential Revision: D56229021
Summary: 4b embedding quantizer Reviewed By: larryliu0820 Differential Revision: D56229021
Summary: 4b embedding quantizer
Differential Revision: D56229021