Open oshrjk opened 1 year ago
I need help with this issue:
(stylegan3) Macbook-de-User:DragGAN User$ python visualizer_drag_gradio.py File under cache_dir (./checkpoints): ['stylegan2-afhqcat-512x512.pkl', 'stylegan2-cat-config-f.pkl', 'stylegan2-car-config-f.pkl', 'stylegan2_lions_512_pytorch.pkl', 'stylegan2_horses_256_pytorch.pkl', 'stylegan2_elephants_512_pytorch.pkl', 'stylegan2-ffhq-512x512.pkl', 'stylegan2_dogs_1024_pytorch.pkl'] Valid checkpoint file: {'stylegan2-afhqcat-512x512': './checkpoints/stylegan2-afhqcat-512x512.pkl', 'stylegan2-cat-config-f': './checkpoints/stylegan2-cat-config-f.pkl', 'stylegan2-car-config-f': './checkpoints/stylegan2-car-config-f.pkl', 'stylegan2_lions_512_pytorch': './checkpoints/stylegan2_lions_512_pytorch.pkl', 'stylegan2_horses_256_pytorch': './checkpoints/stylegan2_horses_256_pytorch.pkl', 'stylegan2_elephants_512_pytorch': './checkpoints/stylegan2_elephants_512_pytorch.pkl', 'stylegan2-ffhq-512x512': './checkpoints/stylegan2-ffhq-512x512.pkl', 'stylegan2_dogs_1024_pytorch': './checkpoints/stylegan2_dogs_1024_pytorch.pkl'} Loading "./checkpoints/stylegan2_lions_512_pytorch.pkl"... Done. () {'z_dim': 512, 'c_dim': 0, 'w_dim': 512, 'img_resolution': 512, 'img_channels': 3, 'mapping_kwargs': {'num_layers': 8, 'embed_features': None, 'layer_features': None, 'activation': 'lrelu', 'lr_multiplier': 0.01, 'w_avg_beta': 0.995}, 'synthesis_kwargs': {'channel_base': 32768, 'channel_max': 512, 'num_fp16_res': 0, 'conv_clamp': None, 'architecture': 'skip', 'resample_filter': [1, 3, 3, 1], 'use_noise': True, 'activation': 'lrelu'}} /Users/User/DragGAN/training/networks_stylegan2.py:264: UserWarning: The operator 'aten::lerp.Scalar_out' is not currently supported on the MPS backend and will fall back to run on the CPU. This may have performance implications. (Triggered internally at /Users/runner/work/_temp/anaconda/conda-bld/pytorch_1682343685278/work/aten/src/ATen/mps/MPSFallback.mm:11.) x = self.w_avg.lerp(x, truncation_psi) Traceback (most recent call last): File "visualizer_drag_gradio.py", line 823, in form_image.select( AttributeError: 'ImageMask' object has no attribute 'select'
Try installing this version:
pip install gradio==3.50.2
I need help with this issue:
(stylegan3) Macbook-de-User:DragGAN User$ python visualizer_drag_gradio.py File under cache_dir (./checkpoints): ['stylegan2-afhqcat-512x512.pkl', 'stylegan2-cat-config-f.pkl', 'stylegan2-car-config-f.pkl', 'stylegan2_lions_512_pytorch.pkl', 'stylegan2_horses_256_pytorch.pkl', 'stylegan2_elephants_512_pytorch.pkl', 'stylegan2-ffhq-512x512.pkl', 'stylegan2_dogs_1024_pytorch.pkl'] Valid checkpoint file: {'stylegan2-afhqcat-512x512': './checkpoints/stylegan2-afhqcat-512x512.pkl', 'stylegan2-cat-config-f': './checkpoints/stylegan2-cat-config-f.pkl', 'stylegan2-car-config-f': './checkpoints/stylegan2-car-config-f.pkl', 'stylegan2_lions_512_pytorch': './checkpoints/stylegan2_lions_512_pytorch.pkl', 'stylegan2_horses_256_pytorch': './checkpoints/stylegan2_horses_256_pytorch.pkl', 'stylegan2_elephants_512_pytorch': './checkpoints/stylegan2_elephants_512_pytorch.pkl', 'stylegan2-ffhq-512x512': './checkpoints/stylegan2-ffhq-512x512.pkl', 'stylegan2_dogs_1024_pytorch': './checkpoints/stylegan2_dogs_1024_pytorch.pkl'} Loading "./checkpoints/stylegan2_lions_512_pytorch.pkl"... Done. () {'z_dim': 512, 'c_dim': 0, 'w_dim': 512, 'img_resolution': 512, 'img_channels': 3, 'mapping_kwargs': {'num_layers': 8, 'embed_features': None, 'layer_features': None, 'activation': 'lrelu', 'lr_multiplier': 0.01, 'w_avg_beta': 0.995}, 'synthesis_kwargs': {'channel_base': 32768, 'channel_max': 512, 'num_fp16_res': 0, 'conv_clamp': None, 'architecture': 'skip', 'resample_filter': [1, 3, 3, 1], 'use_noise': True, 'activation': 'lrelu'}} /Users/User/DragGAN/training/networks_stylegan2.py:264: UserWarning: The operator 'aten::lerp.Scalar_out' is not currently supported on the MPS backend and will fall back to run on the CPU. This may have performance implications. (Triggered internally at /Users/runner/work/_temp/anaconda/conda-bld/pytorch_1682343685278/work/aten/src/ATen/mps/MPSFallback.mm:11.) x = self.w_avg.lerp(x, truncation_psi) Traceback (most recent call last): File "visualizer_drag_gradio.py", line 823, in
form_image.select(
AttributeError: 'ImageMask' object has no attribute 'select'