https://github.com/burhan-q/classifydefectmap_mixedwm38

Classification of wafer defect map patterns
https://github.com/burhan-q/classifydefectmap_mixedwm38

classification wafer-defects wafermap yolov8

Last synced: 2 months ago
JSON representation

Classification of wafer defect map patterns

• Host: GitHub
• URL: https://github.com/burhan-q/classifydefectmap_mixedwm38
• Owner: Burhan-Q
• Created: 2023-05-22T15:16:48.000Z (over 1 year ago)
• Default Branch: master
• Last Pushed: 2023-05-22T16:18:54.000Z (over 1 year ago)
• Last Synced: 2024-10-03T20:41:44.856Z (3 months ago)
• Topics: classification, wafer-defects, wafermap, yolov8
• Language: Python
• Homepage:
• Size: 8.48 MB
• Stars: 6
• Watchers: 1
• Forks: 1
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Wafer Defect Map Classificaiton

This is a demo project built for personal use using the [MixedWM38 dataset](https://github.com/Junliangwangdhu/WaferMap). Note that there is an issue with the dataset as pointed out in [this issue](https://github.com/Junliangwangdhu/WaferMap/issues/2), which was corrected for the results shared here.

## Wafer map patterns

![image](/Wafer%20Map.png)

# Model

Uses the Ultralytics YOLOv8-Large classification model, with standard pretrained weights. The training was run for a short 10 epochs as this was only as a demo project. 

# Results

## Confusion Matrix Result

![image](/wafer_defects/EXP00002/result_confusion_matrix.png)

## Loss and Accuracy Plots

![image](/wafer_defects/EXP00002/results.png)

Overall results from `EXP0002` which was a full GPU training with validation experiment, see the `args.yaml` file to view configuration. Additional metrics were computed using [val_and_results.py](/val_and_results.py). This result should not be considered complete, as model should be trained for additional epochs and additional hyperparameters explored. It is merely a demonstration of implementing a classifier model on wafer defect maps.