Ultra-Light-Fast-Generic-Face-Detector-1MB icon indicating copy to clipboard operation
Ultra-Light-Fast-Generic-Face-Detector-1MB copied to clipboard

Published 20 hours ago • verified publisher icon Linzaer

question about pre-trained model

Open kimtaehyeong opened this issue 4 years ago • 0 comments

Thank you for sharing the good results.

I have a question. What data are the trained models (slim, RFB) provided? Also, I am fine-tuning based on this model. Here, it is labeled (face, background), but I am learning by increasing it to about 5 classes. In conclusion, the performance is poor. I'm guessing it's because the model is too light. Is there a good way? I started learning from Scratch, and there are about 20,000 vocs for face data, and about 20,000 other data. Thank you.

kimtaehyeong avatar Mar 12 '21 01:03 kimtaehyeong