This repository has been archived by the owner on Nov 17, 2023. It is now read-only.
Update imagenet quantization script for MKL-DNN #14407
Closed
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Outdated suggestions cannot be applied.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.
Description
This PR is to skip quantizing FullyConnected layers when its inputs are signed int8 and the subgraph backend is set to MKLDNN, due to the limitation in current version of MKL-DNN. This feature will be enabled when MKL-DNN is upgraded to 0.18 or higher.
For not using MKLDNN subgraph backend (in general mode), users can still use both int8 and uint8 quantized FullyConnected for CPU platform (in this case, int8 will go to IGEMM path, but not providing requantize/dequantize fusion feature).
Checklist
Essentials
Please feel free to remove inapplicable items for your PR.
Changes
Comments
@pengzhao-intel @TaoLv