Megvii-BaseDetection / Megvii-BaseDetection/YOLOX

Performance compared to Yolo4Tiny using TensorRT

Open
#720 5 comments 1 reaction 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
10.6k
Forks
2.5k
PR merge metrics
No merged PRs in 30d

Description

Hello,

We were profiling the performance of YoloX-Tiny by comparing inference time to that of Yolo4-Tiny. This was conducted using a 416 px + FP16 configuration for both models, using TensorRT framework.

Oddly enough, the results shown that Yolo4Tiny is around 40% faster than YoloX-Tiny. This number was similar on multiple machines with different HW. Is this the expected result, or did we possibly make some mistake along the way (e.g. when exporting)?

CUDA Version = 11.1 & 10.2
TensorRT Version = 7.2 & 7.1
HW = Jetson AGX Xavier & RTX 2060

Contributor guide

No contributing guide indexed for this repository

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 reviewing the TensorRT export and benchmarking setup for YOLOX-Tiny and YOLO4-Tiny at 416 px with FP16. Reproduce the comparison on the listed Jetson AGX Xavier or RTX 2060 configurations, then determine whether the roughly 40% performance gap is expected or indicates an export or measurement issue.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning, performance
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
30/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.