This repository has been archived by the owner on Sep 18, 2024. It is now read-only.
Fix dorefa bnn #3247
Merged
Fix dorefa bnn #3247
Conversation
This file contains 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
J-shang
reviewed
Jan 4, 2021
| @@ -148,6 +157,7 @@ def __init__(self, model, config_list, optimizer=None): | |||
| types of nn.module you want to apply quantization, eg. 'Conv2d' | |||
| """ | |||
| super().__init__(model, config_list, optimizer) | |||
| self.quant_grad = QATGrad | |||
There was a problem hiding this comment.
where we use self.quant_grad?
There was a problem hiding this comment.
This is used to clip backward gradient between wmin and wmax.
Cjkkkk
reviewed
Jan 4, 2021
Cjkkkk
reviewed
Jan 4, 2021
Cjkkkk
reviewed
Jan 4, 2021
| 0: "input", | ||
| 1: "weight", | ||
| 2: "output" | ||
| } | ||
|
|
||
| class QuantGrad(torch.autograd.Function): |
There was a problem hiding this comment.
I think the best way to accomplish the quantization gradient calculation is to add quant_weight_grad, quant_input_grad, quant_output_grad method to quantizer class and QuantGrad class only calls this methods instead.
There was a problem hiding this comment.
Agree, this will take us time to discuss a new design and quantizer may be refactored in another pr. In this pr, we only fix dorefa and bnn unfit problem.
J-shang
approved these changes
Jan 4, 2021
Cjkkkk
approved these changes
Jan 4, 2021
QuanluZhang
approved these changes
Jan 4, 2021
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
Because of the modification of compressor to solve the grad error, BNN and Dorefa can't operate normally on current code. This PR fixes this problem.