This repository has been archived by the owner on Nov 17, 2023. It is now read-only.
[Quantization] Support zero-size tensor input for quantization flow #15031
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Outdated suggestions cannot be applied.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 support zero-size tensor input for quantization flow.
Let's take RNN related model as an example, the
begin_state
is always initialized into shape(0, self._num_hidden)
, it worked well in FP32 pass, but failed in INT8 pass due tounknown dimension
error with latest MXNet code base.With this patch, models with such inputs are able to be quantized and run in INT8 mode.
@pengzhao-intel @TaoLv @ZhennanQin
Checklist
Essentials
Please feel free to remove inapplicable items for your PR.
Changes
Comments