3DGAN-Pytorch_Python_下载.zip

# 3DGAN-Pytorch [](https://github.com/meetshah1995/tf-3dgan/blob/master/LICENSE) [](https://arxiv.org/abs/1610.07584) ## Pytorch implementation of 3D Generative Adversarial Network. This is a Pytorch implementation of the paper "Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling". I reference [ tf-3dgan](https://github.com/meetshah1995/tf-3dgan) and translate to pytorch code, Difference is that i use output layer of sigmoid, soft-labelling, learning-rate scheduler. I think all input voxels is positive( 0 or 1), so sigmoid is better than tanh. soft-labelling is good for smoothing loss, I use learning-rate decay to get better quality output ### Requirements * pytoch * scipy * scikit-image ### Usage I use [floydhub](https://www.floydhub.com/) to train model Floydhub is simple deeplearining training tool They offer free-tier for 100h gpu server ``` pip install -U floyd-cli ``` ``` #./input floyd data init chair floyd data upload ``` ``` #./3D_GAN floyd init 3dgan floyd data status floyd run --env pytorch --gpu --data [your data id] "python3 main.py" ``` This porject structure is fitted with floydhub structure, so parent directory contain input, output, 3D_GAN directory ### GAN Trick I use some more trick for better result * the loss function to optimize G is `min (log 1-D)`, but in practice folks practically use `max log D` * Z is Sampled from a gaussian distribution [0, 0.33] * Use Soft Labels - It make loss function smoothing (When I don't use soft labels , I observe divergence after 500 epochs) * learning rate scheduler - after 500 epoch, descriminator's learning rate is decayed If you want to know more trick , go to [Soumith’s ganhacks repo.](https://github.com/soumith/ganhacks) ### Result 1000 epochs <img width="700" height="300" src="./src/293040.png"></img> <img width="1000" height="300" src="./src/293040_visdom.png"></img> ### Reference [ tf-3dgan](https://github.com/meetshah1995/tf-3dgan) [yunjey's pytorch-tutorial](https://github.com/yunjey/pytorch-tutorial)