how to update weights?

[whuyyc]

❔Question

how to update weights for single cls target?

Additional context

for example: I got a weight w1 from trianing set t1 which includes 3000 sonar images of single class. Now I get 10 new images of which the distribution is different from t1 cause the different types of sonar but have the same target class as t1. How should I use the new images to update my weight? Here are 3 methods, but I do not know which one should be chosen:

1. Train the 10 images from w1;
2. Train 3000+10 images from w1;
3. Train 3000+10 images from scratch. For methond 1, I am confused whether the network will learn more about the distribution of new images or not? Because it is similar to train t1 base on weight pretained on COCO. For method 2, does the new weight w2 equal to trianing from n epochs on 10 images and n+m epochs on 3000 images? (w1 is trained well from m epochs on 3000 images)

[github-actions[bot]]

Hello @whuyyc, thank you for your interest in 🚀 YOLOv5! Please visit our ⭐️ Tutorials to get started, where you can find quickstart guides for simple tasks like Custom Data Training all the way to advanced concepts like Hyperparameter Evolution.

If this is a 🐛 Bug Report, please provide screenshots and minimum viable code to reproduce your issue, otherwise we can not help you.

If this is a custom training ❓ Question, please provide as much information as possible, including dataset images, training logs, screenshots, and a public link to online W&B logging if available.

For business inquiries or professional support requests please visit https://www.ultralytics.com or email Glenn Jocher at glenn.jocher@ultralytics.com.

Requirements

Python 3.8 or later with all requirements.txt dependencies installed, including torch>=1.7. To install run:

$ pip install -r requirements.txt

Environments

YOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

• Google Colab Notebook with free GPU:
• Kaggle Notebook with free GPU: https://www.kaggle.com/ultralytics/yolov5
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Docker Image https://hub.docker.com/r/ultralytics/yolov5. See Docker Quickstart Guide

Status

If this badge is green, all YOLOv5 GitHub Actions Continuous Integration (CI) tests are currently passing. CI tests verify correct operation of YOLOv5 training (train.py), testing (test.py), inference (detect.py) and export (export.py) on MacOS, Windows, and Ubuntu every 24 hours and on every commit.

[glenn-jocher]

Use option 2. Option 1 will cause the model to gradually forget the original 3000 images (weight_decay, etc.).

[whuyyc]

Use option 2. Option 1 will cause the model to gradually forget the original 3000 images (weight_decay, etc.).

Thank you very much with worship in my eyes 0-0

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.