[quant][graphmode][fx] Add additional_object_mapping argument to convert #46338
Conversation
Summary: Should we merge quantized module and quantized operator configurations? Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
💊 CI failures summary and remediationsAs of commit fa30163 (more details on the Dr. CI page):
1 failure not recognized by patterns:
This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.Please report bugs/suggestions on the GitHub issue tracker or post in the (internal) Dr. CI Users group. This comment has been revised 3 times. |
💊 CI failures summary and remediationsAs of commit a702a6f (more details on the Dr. CI page): ✅ None of the CI failures appear to be your fault 💚
🚧 1 ongoing upstream failure:These were probably caused by upstream breakages that are not fixed yet:
This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.Please report bugs/suggestions on the GitHub issue tracker or post in the (internal) Dr. CI Users group. This comment has been revised 45 times. |
…ent to convert" Summary: Should we merge quantized module and quantized operator configurations? Test Plan: Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D24317435](https://our.internmc.facebook.com/intern/diff/D24317435) [ghstack-poisoned]
| "static": { | ||
| FloatModule: QuantizedModule, | ||
| float_op: quantized_op | ||
| }, | ||
| "dynamic": { | ||
| FloatModule: DynamicallyQuantizedModule, | ||
| float_op: dynamically_quantized_op | ||
| }, |
There was a problem hiding this comment.
how would custom mappings for QAT, and/or for models which have some parts quantized dynamically and statically fit?
There was a problem hiding this comment.
you mean additional qat module mappings? it's added in the next PR. it's kind of independent from these, it happens before we insert observer, e.g. we swap nn.Conv2d to nn.qat.Conv2d
"static" and "dynamic" here are used during convert, e.g. nn.qat.Conv2d to nn.quantized.Conv2d
…ent to convert" Summary: Should we merge quantized module and quantized operator configurations? Test Plan: Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D24317435](https://our.internmc.facebook.com/intern/diff/D24317435) [ghstack-poisoned]
…ent to convert" Summary: Should we merge quantized module and quantized operator configurations? Test Plan: Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D24317435](https://our.internmc.facebook.com/intern/diff/D24317435) [ghstack-poisoned]
…ent to convert" Summary: Should we merge quantized module and quantized operator configurations? Test Plan: Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D24317435](https://our.internmc.facebook.com/intern/diff/D24317435) [ghstack-poisoned]
…ent to convert" Summary: Should we merge quantized module and quantized operator configurations? Test Plan: Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D24317435](https://our.internmc.facebook.com/intern/diff/D24317435) [ghstack-poisoned]
…ent to convert" Summary: Should we merge quantized module and quantized operator configurations? Test Plan: Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D24317435](https://our.internmc.facebook.com/intern/diff/D24317435) [ghstack-poisoned]
|
This pull request has been merged in 746febd. |
Stack from ghstack:
Summary:
Should we merge quantized module and quantized operator configurations?
Test Plan:
Reviewers:
Subscribers:
Tasks:
Tags:
Differential Revision: D24317435