NVIDIA / NVIDIA/TensorRT

INT8 Quantization of dinov2 TensorRT Model is Not Faster than FP16 Quantization

Open
#4,273 2 comments 0 reactions 1 assignee View on GitHub

@akhilg-nv is already working on this.

Since Dec 16, 2024.

Module:Quantization triaged
Dominant language
C++
Stars
13.4k
Forks
2.4k
Avg merge
5d 3h
Merged PRs (30d)
2

Description

Hello,

I used PyTorch-Quantization for post-training INT8 quantization on the dinov2-base model and then converted it to a TensorRT model. However, I found that the INT8 model is slightly slower than the FP16 model (the same conclusion was observed on A100, V100, and A10). Is this behavior normal?

Thank you.

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.