怎样去训练一个1.4版本的GFPGAN #356

New Issue

Closed

opened 2026-01-29 21:47:06 +00:00 by claunia · 1 comment

claunia commented 2026-01-29 21:47:06 +00:00

Owner

Copy Link

Originally created by @Happyrealone on GitHub (Jun 12, 2023).

我尝试训练一个新的模型，将BasicSR中train_StyleGAN2_256_Cmul2_FFHQ.yml的参数文件改成了out_size为512，channel_multiplier为1的版本，并成功训练出了一个StyleGAN2的生成器，然后我通过同名替换的方式将1.4版本的GFPGAN模型参数改成了训练好的StyleGAN2生成器的参数，但是在后续应用中出现了数据尺寸不匹配的问题，应该怎样处理呢？
Error(s) in loading state_dict for GFPGANv1Clean:
size mismatch for stylegan_decoder.style_convs.6.modulated_conv.weight: copying a param with shape torch.Size([1, 256, 512, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 512, 512, 3, 3]).
size mismatch for stylegan_decoder.style_convs.7.modulated_conv.weight: copying a param with shape torch.Size([1, 256, 256, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 512, 512, 3, 3]).
size mismatch for stylegan_decoder.style_convs.7.modulated_conv.modulation.weight: copying a param with shape torch.Size([256, 512]) from checkpoint, the shape in current model is torch.Size([512, 512]).
size mismatch for stylegan_decoder.style_convs.7.modulated_conv.modulation.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([512]).
size mismatch for stylegan_decoder.style_convs.8.modulated_conv.weight: copying a param with shape torch.Size([1, 128, 256, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 256, 512, 3, 3]).
size mismatch for stylegan_decoder.style_convs.8.modulated_conv.modulation.weight: copying a param with shape torch.Size([256, 512]) from checkpoint, the shape in current model is torch.Size([512, 512]).
size mismatch for stylegan_decoder.style_convs.8.modulated_conv.modulation.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([512]).
size mismatch for stylegan_decoder.style_convs.9.modulated_conv.weight: copying a param with shape torch.Size([1, 128, 128, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 256, 256, 3, 3]).
size mismatch for stylegan_decoder.style_convs.9.modulated_conv.modulation.weight: copying a param with shape torch.Size([128, 512]) from checkpoint, the shape in current model is torch.Size([256, 512]).
size mismatch for stylegan_decoder.style_convs.9.modulated_conv.modulation.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for stylegan_decoder.style_convs.10.modulated_conv.weight: copying a param with shape torch.Size([1, 64, 128, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 128, 256, 3, 3]).
size mismatch for stylegan_decoder.style_convs.10.modulated_conv.modulation.weight: copying a param with shape torch.Size([128, 512]) from checkpoint, the shape in current model is torch.Size([256, 512]).
size mismatch for stylegan_decoder.style_convs.10.modulated_conv.modulation.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for stylegan_decoder.style_convs.11.modulated_conv.weight: copying a param with shape torch.Size([1, 64, 64, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 128, 128, 3, 3]).
size mismatch for stylegan_decoder.style_convs.11.modulated_conv.modulation.weight: copying a param with shape torch.Size([64, 512]) from checkpoint, the shape in current model is torch.Size([128, 512]).
size mismatch for stylegan_decoder.style_convs.11.modulated_conv.modulation.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for stylegan_decoder.style_convs.12.modulated_conv.weight: copying a param with shape torch.Size([1, 32, 64, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 64, 128, 3, 3]).
size mismatch for stylegan_decoder.style_convs.12.modulated_conv.modulation.weight: copying a param with shape torch.Size([64, 512]) from checkpoint, the shape in current model is torch.Size([128, 512]).
size mismatch for stylegan_decoder.style_convs.12.modulated_conv.modulation.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for stylegan_decoder.style_convs.13.modulated_conv.weight: copying a param with shape torch.Size([1, 32, 32, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 64, 64, 3, 3]).
size mismatch for stylegan_decoder.style_convs.13.modulated_conv.modulation.weight: copying a param with shape torch.Size([32, 512]) from checkpoint, the shape in current model is torch.Size([64, 512]).
size mismatch for stylegan_decoder.style_convs.13.modulated_conv.modulation.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for stylegan_decoder.to_rgbs.3.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 512, 1, 1]).
size mismatch for stylegan_decoder.to_rgbs.3.modulated_conv.modulation.weight: copying a param with shape torch.Size([256, 512]) from checkpoint, the shape in current model is torch.Size([512, 512]).
size mismatch for stylegan_decoder.to_rgbs.3.modulated_conv.modulation.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([512]).
size mismatch for stylegan_decoder.to_rgbs.4.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 256, 1, 1]).
size mismatch for stylegan_decoder.to_rgbs.4.modulated_conv.modulation.weight: copying a param with shape torch.Size([128, 512]) from checkpoint, the shape in current model is torch.Size([256, 512]).
size mismatch for stylegan_decoder.to_rgbs.4.modulated_conv.modulation.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for stylegan_decoder.to_rgbs.5.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 128, 1, 1]).
size mismatch for stylegan_decoder.to_rgbs.5.modulated_conv.modulation.weight: copying a param with shape torch.Size([64, 512]) from checkpoint, the shape in current model is torch.Size([128, 512]).
size mismatch for stylegan_decoder.to_rgbs.5.modulated_conv.modulation.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for stylegan_decoder.to_rgbs.6.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 32, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 64, 1, 1]).
size mismatch for stylegan_decoder.to_rgbs.6.modulated_conv.modulation.weight: copying a param with shape torch.Size([32, 512]) from checkpoint, the shape in current model is torch.Size([64, 512]).
size mismatch for stylegan_decoder.to_rgbs.6.modulated_conv.modulation.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
File "/home/pt/miniconda3/envs/sadtalker/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1604, in load_state_dict
raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
File "/home/pt/miniconda3/envs/sadtalker/lib/python3.8/site-packages/gfpgan/utils.py", line 99, in init
self.gfpgan.load_state_dict(loadnet[keyname], strict=True)
File "/home/pt/Project/SadTalker/src/utils/face_enhancer.py", line 131, in enhancer
restorer = GFPGANer(
File "/home/pt/Project/SadTalker/src/facerender/animate.py", line 210, in generate
enhanced_images = face_enhancer(full_video_path, method=enhancer, bg_upsampler=background_enhancer) # type: ignore
File "/home/pt/Project/SadTalker/inference.py", line 113, in main
animate_from_coeff.generate(data, save_dir, pic_path, crop_info,
File "/home/pt/Project/SadTalker/inference.py", line 160, in
main(args)
File "/home/pt/miniconda3/envs/sadtalker/lib/python3.8/runpy.py", line 87, in _run_code
exec(code, run_globals)
File "/home/pt/miniconda3/envs/sadtalker/lib/python3.8/runpy.py", line 194, in _run_module_as_main (Current frame)
return _run_code(code, main_globals, None,
RuntimeError: Error(s) in loading state_dict for GFPGANv1Clean:
size mismatch for stylegan_decoder.style_convs.6.modulated_conv.weight: copying a param with shape torch.Size([1, 256, 512, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 512, 512, 3, 3]).
size mismatch for stylegan_decoder.style_convs.7.modulated_conv.weight: copying a param with shape torch.Size([1, 256, 256, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 512, 512, 3, 3]).
size mismatch for stylegan_decoder.style_convs.7.modulated_conv.modulation.weight: copying a param with shape torch.Size([256, 512]) from checkpoint, the shape in current model is torch.Size([512, 512]).
size mismatch for stylegan_decoder.style_convs.7.modulated_conv.modulation.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([512]).
size mismatch for stylegan_decoder.style_convs.8.modulated_conv.weight: copying a param with shape torch.Size([1, 128, 256, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 256, 512, 3, 3]).
size mismatch for stylegan_decoder.style_convs.8.modulated_conv.modulation.weight: copying a param with shape torch.Size([256, 512]) from checkpoint, the shape in current model is torch.Size([512, 512]).
size mismatch for stylegan_decoder.style_convs.8.modulated_conv.modulation.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([512]).
size mismatch for stylegan_decoder.style_convs.9.modulated_conv.weight: copying a param with shape torch.Size([1, 128, 128, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 256, 256, 3, 3]).
size mismatch for stylegan_decoder.style_convs.9.modulated_conv.modulation.weight: copying a param with shape torch.Size([128, 512]) from checkpoint, the shape in current model is torch.Size([256, 512]).
size mismatch for stylegan_decoder.style_convs.9.modulated_conv.modulation.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for stylegan_decoder.style_convs.10.modulated_conv.weight: copying a param with shape torch.Size([1, 64, 128, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 128, 256, 3, 3]).
size mismatch for stylegan_decoder.style_convs.10.modulated_conv.modulation.weight: copying a param with shape torch.Size([128, 512]) from checkpoint, the shape in current model is torch.Size([256, 512]).
size mismatch for stylegan_decoder.style_convs.10.modulated_conv.modulation.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for stylegan_decoder.style_convs.11.modulated_conv.weight: copying a param with shape torch.Size([1, 64, 64, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 128, 128, 3, 3]).
size mismatch for stylegan_decoder.style_convs.11.modulated_conv.modulation.weight: copying a param with shape torch.Size([64, 512]) from checkpoint, the shape in current model is torch.Size([128, 512]).
size mismatch for stylegan_decoder.style_convs.11.modulated_conv.modulation.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for stylegan_decoder.style_convs.12.modulated_conv.weight: copying a param with shape torch.Size([1, 32, 64, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 64, 128, 3, 3]).
size mismatch for stylegan_decoder.style_convs.12.modulated_conv.modulation.weight: copying a param with shape torch.Size([64, 512]) from checkpoint, the shape in current model is torch.Size([128, 512]).
size mismatch for stylegan_decoder.style_convs.12.modulated_conv.modulation.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for stylegan_decoder.style_convs.13.modulated_conv.weight: copying a param with shape torch.Size([1, 32, 32, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 64, 64, 3, 3]).
size mismatch for stylegan_decoder.style_convs.13.modulated_conv.modulation.weight: copying a param with shape torch.Size([32, 512]) from checkpoint, the shape in current model is torch.Size([64, 512]).
size mismatch for stylegan_decoder.style_convs.13.modulated_conv.modulation.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).
size mismatch for stylegan_decoder.to_rgbs.3.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 512, 1, 1]).
size mismatch for stylegan_decoder.to_rgbs.3.modulated_conv.modulation.weight: copying a param with shape torch.Size([256, 512]) from checkpoint, the shape in current model is torch.Size([512, 512]).
size mismatch for stylegan_decoder.to_rgbs.3.modulated_conv.modulation.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([512]).
size mismatch for stylegan_decoder.to_rgbs.4.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 256, 1, 1]).
size mismatch for stylegan_decoder.to_rgbs.4.modulated_conv.modulation.weight: copying a param with shape torch.Size([128, 512]) from checkpoint, the shape in current model is torch.Size([256, 512]).
size mismatch for stylegan_decoder.to_rgbs.4.modulated_conv.modulation.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]).
size mismatch for stylegan_decoder.to_rgbs.5.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 128, 1, 1]).
size mismatch for stylegan_decoder.to_rgbs.5.modulated_conv.modulation.weight: copying a param with shape torch.Size([64, 512]) from checkpoint, the shape in current model is torch.Size([128, 512]).
size mismatch for stylegan_decoder.to_rgbs.5.modulated_conv.modulation.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for stylegan_decoder.to_rgbs.6.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 32, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 64, 1, 1]).
size mismatch for stylegan_decoder.to_rgbs.6.modulated_conv.modulation.weight: copying a param with shape torch.Size([32, 512]) from checkpoint, the shape in current model is torch.Size([64, 512]).
size mismatch for stylegan_decoder.to_rgbs.6.modulated_conv.modulation.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).

Originally created by @Happyrealone on GitHub (Jun 12, 2023). 我尝试训练一个新的模型，将BasicSR中train_StyleGAN2_256_Cmul2_FFHQ.yml的参数文件改成了out_size为512，channel_multiplier为1的版本，并成功训练出了一个StyleGAN2的生成器，然后我通过同名替换的方式将1.4版本的GFPGAN模型参数改成了训练好的StyleGAN2生成器的参数，但是在后续应用中出现了数据尺寸不匹配的问题，应该怎样处理呢？ Error(s) in loading state_dict for GFPGANv1Clean: size mismatch for stylegan_decoder.style_convs.6.modulated_conv.weight: copying a param with shape torch.Size([1, 256, 512, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 512, 512, 3, 3]). size mismatch for stylegan_decoder.style_convs.7.modulated_conv.weight: copying a param with shape torch.Size([1, 256, 256, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 512, 512, 3, 3]). size mismatch for stylegan_decoder.style_convs.7.modulated_conv.modulation.weight: copying a param with shape torch.Size([256, 512]) from checkpoint, the shape in current model is torch.Size([512, 512]). size mismatch for stylegan_decoder.style_convs.7.modulated_conv.modulation.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([512]). size mismatch for stylegan_decoder.style_convs.8.modulated_conv.weight: copying a param with shape torch.Size([1, 128, 256, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 256, 512, 3, 3]). size mismatch for stylegan_decoder.style_convs.8.modulated_conv.modulation.weight: copying a param with shape torch.Size([256, 512]) from checkpoint, the shape in current model is torch.Size([512, 512]). size mismatch for stylegan_decoder.style_convs.8.modulated_conv.modulation.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([512]). size mismatch for stylegan_decoder.style_convs.9.modulated_conv.weight: copying a param with shape torch.Size([1, 128, 128, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 256, 256, 3, 3]). size mismatch for stylegan_decoder.style_convs.9.modulated_conv.modulation.weight: copying a param with shape torch.Size([128, 512]) from checkpoint, the shape in current model is torch.Size([256, 512]). size mismatch for stylegan_decoder.style_convs.9.modulated_conv.modulation.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]). size mismatch for stylegan_decoder.style_convs.10.modulated_conv.weight: copying a param with shape torch.Size([1, 64, 128, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 128, 256, 3, 3]). size mismatch for stylegan_decoder.style_convs.10.modulated_conv.modulation.weight: copying a param with shape torch.Size([128, 512]) from checkpoint, the shape in current model is torch.Size([256, 512]). size mismatch for stylegan_decoder.style_convs.10.modulated_conv.modulation.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]). size mismatch for stylegan_decoder.style_convs.11.modulated_conv.weight: copying a param with shape torch.Size([1, 64, 64, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 128, 128, 3, 3]). size mismatch for stylegan_decoder.style_convs.11.modulated_conv.modulation.weight: copying a param with shape torch.Size([64, 512]) from checkpoint, the shape in current model is torch.Size([128, 512]). size mismatch for stylegan_decoder.style_convs.11.modulated_conv.modulation.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]). size mismatch for stylegan_decoder.style_convs.12.modulated_conv.weight: copying a param with shape torch.Size([1, 32, 64, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 64, 128, 3, 3]). size mismatch for stylegan_decoder.style_convs.12.modulated_conv.modulation.weight: copying a param with shape torch.Size([64, 512]) from checkpoint, the shape in current model is torch.Size([128, 512]). size mismatch for stylegan_decoder.style_convs.12.modulated_conv.modulation.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]). size mismatch for stylegan_decoder.style_convs.13.modulated_conv.weight: copying a param with shape torch.Size([1, 32, 32, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 64, 64, 3, 3]). size mismatch for stylegan_decoder.style_convs.13.modulated_conv.modulation.weight: copying a param with shape torch.Size([32, 512]) from checkpoint, the shape in current model is torch.Size([64, 512]). size mismatch for stylegan_decoder.style_convs.13.modulated_conv.modulation.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]). size mismatch for stylegan_decoder.to_rgbs.3.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 512, 1, 1]). size mismatch for stylegan_decoder.to_rgbs.3.modulated_conv.modulation.weight: copying a param with shape torch.Size([256, 512]) from checkpoint, the shape in current model is torch.Size([512, 512]). size mismatch for stylegan_decoder.to_rgbs.3.modulated_conv.modulation.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([512]). size mismatch for stylegan_decoder.to_rgbs.4.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 256, 1, 1]). size mismatch for stylegan_decoder.to_rgbs.4.modulated_conv.modulation.weight: copying a param with shape torch.Size([128, 512]) from checkpoint, the shape in current model is torch.Size([256, 512]). size mismatch for stylegan_decoder.to_rgbs.4.modulated_conv.modulation.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]). size mismatch for stylegan_decoder.to_rgbs.5.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 128, 1, 1]). size mismatch for stylegan_decoder.to_rgbs.5.modulated_conv.modulation.weight: copying a param with shape torch.Size([64, 512]) from checkpoint, the shape in current model is torch.Size([128, 512]). size mismatch for stylegan_decoder.to_rgbs.5.modulated_conv.modulation.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]). size mismatch for stylegan_decoder.to_rgbs.6.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 32, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 64, 1, 1]). size mismatch for stylegan_decoder.to_rgbs.6.modulated_conv.modulation.weight: copying a param with shape torch.Size([32, 512]) from checkpoint, the shape in current model is torch.Size([64, 512]). size mismatch for stylegan_decoder.to_rgbs.6.modulated_conv.modulation.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]). File "/home/pt/miniconda3/envs/sadtalker/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1604, in load_state_dict raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format( File "/home/pt/miniconda3/envs/sadtalker/lib/python3.8/site-packages/gfpgan/utils.py", line 99, in __init__ self.gfpgan.load_state_dict(loadnet[keyname], strict=True) File "/home/pt/Project/SadTalker/src/utils/face_enhancer.py", line 131, in enhancer restorer = GFPGANer( File "/home/pt/Project/SadTalker/src/facerender/animate.py", line 210, in generate enhanced_images = face_enhancer(full_video_path, method=enhancer, bg_upsampler=background_enhancer) # type: ignore File "/home/pt/Project/SadTalker/inference.py", line 113, in main animate_from_coeff.generate(data, save_dir, pic_path, crop_info, \ File "/home/pt/Project/SadTalker/inference.py", line 160, in <module> main(args) File "/home/pt/miniconda3/envs/sadtalker/lib/python3.8/runpy.py", line 87, in _run_code exec(code, run_globals) File "/home/pt/miniconda3/envs/sadtalker/lib/python3.8/runpy.py", line 194, in _run_module_as_main (Current frame) return _run_code(code, main_globals, None, RuntimeError: Error(s) in loading state_dict for GFPGANv1Clean: size mismatch for stylegan_decoder.style_convs.6.modulated_conv.weight: copying a param with shape torch.Size([1, 256, 512, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 512, 512, 3, 3]). size mismatch for stylegan_decoder.style_convs.7.modulated_conv.weight: copying a param with shape torch.Size([1, 256, 256, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 512, 512, 3, 3]). size mismatch for stylegan_decoder.style_convs.7.modulated_conv.modulation.weight: copying a param with shape torch.Size([256, 512]) from checkpoint, the shape in current model is torch.Size([512, 512]). size mismatch for stylegan_decoder.style_convs.7.modulated_conv.modulation.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([512]). size mismatch for stylegan_decoder.style_convs.8.modulated_conv.weight: copying a param with shape torch.Size([1, 128, 256, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 256, 512, 3, 3]). size mismatch for stylegan_decoder.style_convs.8.modulated_conv.modulation.weight: copying a param with shape torch.Size([256, 512]) from checkpoint, the shape in current model is torch.Size([512, 512]). size mismatch for stylegan_decoder.style_convs.8.modulated_conv.modulation.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([512]). size mismatch for stylegan_decoder.style_convs.9.modulated_conv.weight: copying a param with shape torch.Size([1, 128, 128, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 256, 256, 3, 3]). size mismatch for stylegan_decoder.style_convs.9.modulated_conv.modulation.weight: copying a param with shape torch.Size([128, 512]) from checkpoint, the shape in current model is torch.Size([256, 512]). size mismatch for stylegan_decoder.style_convs.9.modulated_conv.modulation.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]). size mismatch for stylegan_decoder.style_convs.10.modulated_conv.weight: copying a param with shape torch.Size([1, 64, 128, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 128, 256, 3, 3]). size mismatch for stylegan_decoder.style_convs.10.modulated_conv.modulation.weight: copying a param with shape torch.Size([128, 512]) from checkpoint, the shape in current model is torch.Size([256, 512]). size mismatch for stylegan_decoder.style_convs.10.modulated_conv.modulation.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]). size mismatch for stylegan_decoder.style_convs.11.modulated_conv.weight: copying a param with shape torch.Size([1, 64, 64, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 128, 128, 3, 3]). size mismatch for stylegan_decoder.style_convs.11.modulated_conv.modulation.weight: copying a param with shape torch.Size([64, 512]) from checkpoint, the shape in current model is torch.Size([128, 512]). size mismatch for stylegan_decoder.style_convs.11.modulated_conv.modulation.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]). size mismatch for stylegan_decoder.style_convs.12.modulated_conv.weight: copying a param with shape torch.Size([1, 32, 64, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 64, 128, 3, 3]). size mismatch for stylegan_decoder.style_convs.12.modulated_conv.modulation.weight: copying a param with shape torch.Size([64, 512]) from checkpoint, the shape in current model is torch.Size([128, 512]). size mismatch for stylegan_decoder.style_convs.12.modulated_conv.modulation.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]). size mismatch for stylegan_decoder.style_convs.13.modulated_conv.weight: copying a param with shape torch.Size([1, 32, 32, 3, 3]) from checkpoint, the shape in current model is torch.Size([1, 64, 64, 3, 3]). size mismatch for stylegan_decoder.style_convs.13.modulated_conv.modulation.weight: copying a param with shape torch.Size([32, 512]) from checkpoint, the shape in current model is torch.Size([64, 512]). size mismatch for stylegan_decoder.style_convs.13.modulated_conv.modulation.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]). size mismatch for stylegan_decoder.to_rgbs.3.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 512, 1, 1]). size mismatch for stylegan_decoder.to_rgbs.3.modulated_conv.modulation.weight: copying a param with shape torch.Size([256, 512]) from checkpoint, the shape in current model is torch.Size([512, 512]). size mismatch for stylegan_decoder.to_rgbs.3.modulated_conv.modulation.bias: copying a param with shape torch.Size([256]) from checkpoint, the shape in current model is torch.Size([512]). size mismatch for stylegan_decoder.to_rgbs.4.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 256, 1, 1]). size mismatch for stylegan_decoder.to_rgbs.4.modulated_conv.modulation.weight: copying a param with shape torch.Size([128, 512]) from checkpoint, the shape in current model is torch.Size([256, 512]). size mismatch for stylegan_decoder.to_rgbs.4.modulated_conv.modulation.bias: copying a param with shape torch.Size([128]) from checkpoint, the shape in current model is torch.Size([256]). size mismatch for stylegan_decoder.to_rgbs.5.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 128, 1, 1]). size mismatch for stylegan_decoder.to_rgbs.5.modulated_conv.modulation.weight: copying a param with shape torch.Size([64, 512]) from checkpoint, the shape in current model is torch.Size([128, 512]). size mismatch for stylegan_decoder.to_rgbs.5.modulated_conv.modulation.bias: copying a param with shape torch.Size([64]) from checkpoint, the shape in current model is torch.Size([128]). size mismatch for stylegan_decoder.to_rgbs.6.modulated_conv.weight: copying a param with shape torch.Size([1, 3, 32, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 3, 64, 1, 1]). size mismatch for stylegan_decoder.to_rgbs.6.modulated_conv.modulation.weight: copying a param with shape torch.Size([32, 512]) from checkpoint, the shape in current model is torch.Size([64, 512]). size mismatch for stylegan_decoder.to_rgbs.6.modulated_conv.modulation.bias: copying a param with shape torch.Size([32]) from checkpoint, the shape in current model is torch.Size([64]).

claunia closed this issue 2026-01-29 21:47:07 +00:00

claunia commented 2026-01-29 21:47:07 +00:00

Author

Owner

Copy Link

@MOr1s2 commented on GitHub (Jul 1, 2024):

尺寸不匹配是需要训练channel_multiplier=2的styleganv2，请问你现在有成功训练出来GFPGANv1.4的结果吗？

@MOr1s2 commented on GitHub (Jul 1, 2024): 尺寸不匹配是需要训练channel_multiplier=2的styleganv2，请问你现在有成功训练出来GFPGANv1.4的结果吗？

claunia referenced this issue 2026-01-29 21:48:03 +00:00

How to train on 1024 x 1024 face images? #475

Sign in to join this conversation.

No Branch/Tag Specified

Branches Tags

master

v1.3.8

v1.3.7

v1.3.6

v1.3.5

v1.3.4

v1.3.3

v1.3.2

v1.3.1

v1.3.0

v0.2.4

v0.2.3

v0.2.2

v.0.2.1

v0.2.0

v0.1.0

Labels

Clear labels

pull-request

Mirrored from GitHub Pull Request

No Label

Milestone

No items

No Milestone

Projects

Clear projects

No project

Assignees

Clear assignees

claunia

No Assignees

1 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: TencentARC/GFPGAN#356