tensorflow / tensorflow/model-optimization

Is there a way to do symmetric quantization for activations in post-training quantization

Open
#801 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
1.6k
Forks
349
Avg merge
3d 2h
Merged PRs (30d)
1

Description

https://tensorflow.google.cn/lite/performance/quantization_spec
from this document we can see the post-training quantization do asymmetric quantization for activations, do symmetric quantization for wieghts.

I follow this guide: https://tensorflow.google.cn/lite/performance/post_training_integer_quant do full int8 quantization for my network mode is OK, as the above said the weights is symmetric , the feature is asymmetric.

My question is I also want the activations also is symmetric, How to do that?

thanks!

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reading the linked post-training quantization guide and quantization specification, then inspect the model-optimization quantization entry points for activation configuration. The issue does not name files or tests; completion would require establishing whether symmetric activation quantization is supported and documenting or implementing a defined path for it.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, tensorflow
Domain
machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.