James Fury

Where does the model to get the option from terminal

Out of memory

2

Why the program can run in the training, but stuck in saving weight? Saving weights to /data1/code/BigGAN-PyTorch/weights/BigGAN_I512_seed0_Gch96_Dch96_bs8_nDa8_nGa8_Glr1.0e-04_Dlr4.0e-04_Gnlinplace_relu_Dnlinplace_relu_Ginitortho_Dinitortho_Gattn64_Dattn64_Gshared_hier_ema... Saving weights to /data1/code/BigGAN-PyTorch/weights/BigGAN_I512_seed0_Gch96_Dch96_bs8_nDa8_nGa8_Glr1.0e-04_Dlr4.0e-04_Gnlinplace_relu_Dnlinplace_relu_Ginitortho_Dinitortho_Gattn64_Dattn64_Gshared_hier_ema/copy0... RuntimeError: CUDA out of memory. Tried to allocate 3.75...

how to solve

About Cls_cvt.py

2

I want to konw why there isn't any to the function get_cls_model and compute_macs

Equal linear

1

Hi, excuse me. What's the exact meaning of the Equalinear? Why need to scale the Weight and bias

[Code]Is there any script for evaluation?

1

Hi, I see the table 1 in your paper. Is there any evaluation script provided for this ?

Thanks for your beautiful work, i want to konw how to do the interpolation experiment just like the result in the paper specifically ?

Is there any evaluation script for SSIM metrics?

issure about paper

1

您好，我想请教下论文关于同时控制多个属性不变而对一个属性进行操作，怎么能找到一个同时保持多个属性不变的分界面呢？ 每个属性都有一个分界面，多个属性的分界面并不会重合啊？ 而且论文对应的那个实验图也没有特别看懂，可以稍微解释下吗？[  ](url)

Hi, the code to realize the flow `flow[0,:,:,0] = cv2.resize(np.float32(mvCont[0,:,:,1]), (0,0), fx=feat_size[1]/input_size[1], fy=feat_size[0]/input_size[0], interpolation = cv2.INTER_LINEAR)*feat_size[0]/input_size[0]` `flow[0,:,:,1] = cv2.resize(np.float32(mvCont[0,:,:,0]), (0,0), fx=feat_size[1]/input_size[1], fy=feat_size[0]/input_size[0], interpolation = cv2.INTER_LINEAR)*feat_size[1]/input_size[1]` Why there is reversion that...