Training loss does not change

agrimgupta92 / sgan

Code for "Social GAN: Socially Acceptable Trajectories with Generative Adversarial Networks", Gupta et al, CVPR 2018

MIT License

819 stars 260 forks source link

Training loss does not change #2

Closed tessavdheiden closed 6 years ago

tessavdheiden commented 6 years ago

Hi Agrim,

The description of the code is super and working with the pretrained models as well!

However, I am trying to train it myself, but the generator's and discriminator's loss do not drop. Can you please let me know if the loss changed for your setting?

Best Tessa

agrimgupta92 commented 6 years ago

Try using print args scripts on one of the pre trained models and then use the same hyper parameter for your training. I would also suggest looking at weight initializations they helped with training.

agoodge commented 5 years ago

Hi Tessa,

I'm having the same problem now. I tried using the same hyper parameters as for the zara1_12 model, however it didn't help. Do you remember how you fixed the issue?

Many thanks