mlcommons / mlcommons/mobile_models

MobileBERT tflite int8 model seems not follow quantization spec

Open
#21 6 comments 0 reactions 2 assignees View on GitHub

@jwookiehong is already working on this.

Since Jun 21, 2021.

Dominant language
No language data
Stars
26
Forks
10
No merged PRs in 30d

Description

The model downloaded from https://github.com/fatihcakirs/mobile_models/blob/main/v0_7/tflite/mobilebert_int8_384_20200602.tflite

Some Fully-connected weights has none-zero zero point (ex. weight bert/encoder/layer_0/attention/self/MatMul19 has zero-point = 6) , which violate the TFLite quantization spec.

I am afraid this might cause issues on some implementation which bypass the FC weight zero-point calculation.

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.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.