Re-Trainning Model #136
Description
Hello,
First, Thanks for sharing this project code.
Before implementing my model, I am training your code now.
I have already generate the uv map data from 300W-LP(122450 samples)
and modify .#L39 line like this
Except that I did not change the other parts (learning_rate, epochs, batch_size).
But the results are still do not look like the original version yet.
I cannot sure this is because of 'wrong preprocessing the uv map', 'little training steps' or 'wrong modification on the code'.
How did you set your training code? Could I have your advices?
Upload the final model test image and the training loss from train.py
