Training Accuracy stops improving

yu4u / cutout-random-erasing

Cutout / Random Erasing implementation, especially for ImageDataGenerator in Keras

MIT License

167 stars 38 forks source link

Open shuaibaslam2019 opened 3 years ago

shuaibaslam2019 commented 3 years ago

When I apply your purposed data augmentation technique, it stops my training accuracy to 95.90 ? What Could be the reason?

preprocessing_function=get_random_eraser(p=0.8, v_l=0, v_h=1, pixel_level=False))

Sicily-F commented 3 years ago

can i ask, did you have issues performing the pixel randomisation? Or did you use the grey/black squares for the random erasing

Sicily-F commented 3 years ago

hi again- any chance you could check out my question here: https://github.com/yu4u/cutout-random-erasing/issues/6 I've still not had a reply from the owner of this repo

shuaibaslam2019 commented 3 years ago

Hi @Sicily-F ,

Check my updated code in the relevent thread.

Best regards,