roboflow / roboflow/roboflow-python

Yolov9 bad results

Open
#332 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
629
Forks
140
Avg merge
2d 2h
Merged PRs (30d)
5

Description

I want to train a model with personal protective equipment. When I train with a helmet, vest and mask, the results are good, but when I add gloves and goggles in addition to these equipment by taking images from different datasets, the results get quite bad. Is there a dataset or approach you can recommend?

I need a dataset of situations with and without hardhats, vests, masks, gloves and goggles(Hardhat,No-Hardhat, Glove, No-Glove etc.)

https://universe.roboflow.com/yavuzhan-zom9x/mergeddataset-50xku/dataset/19 (Dataset I created by merging)
https://universe.roboflow.com/roboflow-universe-projects/construction-site-safety (The dataset I use for hardhats, vests and masks)

image
I am presenting the success evaluation metrics of the last model I trained in the screenshot.

Thanks for your support.

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

No repository file, test, or entry point is mentioned. Start by comparing the linked merged dataset with the construction-site-safety dataset and the reported evaluation metrics, focusing on label coverage and dataset composition. Done means documenting whether the data explains the degraded results and identifying a suitable dataset or training approach.

Written by the indexing model from the issue text.

Assessment

Tech stack
computer-vision, machine-learning, python
Domain
computer-vision, machine-learning
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
15/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.