转换模型报错，如何定位

XdpAreKid commented 6 months ago

error log | 日志或报错信息 | ログ

./pnnx gfpganv1_clean_model.pt inputshape=[1,3,512,512] fp16=0 pnnxparam = gfpganv1_clean_model.pnnx.param pnnxbin = gfpganv1_clean_model.pnnx.bin pnnxpy = gfpganv1_clean_model_pnnx.py pnnxonnx = gfpganv1_clean_model.pnnx.onnx ncnnparam = gfpganv1_clean_model.ncnn.param ncnnbin = gfpganv1_clean_model.ncnn.bin ncnnpy = gfpganv1_clean_model_ncnn.py fp16 = 0 optlevel = 2 device = cpu inputshape = [1,3,512,512]f32 inputshape2 = customop = moduleop = ############# pass_level0 inline module = gfpgan.archs.gfpganv1_clean_arch.ResBlock inline module = gfpgan.archs.gfpganv1_clean_arch.StyleGAN2GeneratorCSFT inline module = gfpgan.archs.stylegan2_clean_arch.ConstantInput inline module = gfpgan.archs.stylegan2_clean_arch.ModulatedConv2d inline module = gfpgan.archs.stylegan2_clean_arch.StyleConv inline module = gfpgan.archs.stylegan2_clean_arch.ToRGB inline module = gfpgan.archs.gfpganv1_clean_arch.ResBlock inline module = gfpgan.archs.gfpganv1_clean_arch.StyleGAN2GeneratorCSFT inline module = gfpgan.archs.stylegan2_clean_arch.ConstantInput inline module = gfpgan.archs.stylegan2_clean_arch.ModulatedConv2d inline module = gfpgan.archs.stylegan2_clean_arch.StyleConv inline module = gfpgan.archs.stylegan2_clean_arch.ToRGB

inline module = gfpgan.archs.gfpganv1_clean_arch.ResBlock inline module = gfpgan.archs.gfpganv1_clean_arch.StyleGAN2GeneratorCSFT inline module = gfpgan.archs.stylegan2_clean_arch.ConstantInput inline module = gfpgan.archs.stylegan2_clean_arch.ModulatedConv2d inline module = gfpgan.archs.stylegan2_clean_arch.StyleConv inline module = gfpgan.archs.stylegan2_clean_arch.ToRGB

############# pass_level1 ############# pass_level2 ############# pass_level3 ############# pass_level4 ############# pass_level5 ############# pass_ncnn fallback batch axis 233 for operand 196 fallback batch axis 233 for operand 207 fallback batch axis 233 for operand 221 fallback batch axis 233 for operand 243 fallback batch axis 233 for operand 254 fallback batch axis 233 for operand 269 fallback batch axis 233 for operand 291 fallback batch axis 233 for operand 302 fallback batch axis 233 for operand 317 fallback batch axis 233 for operand 339 fallback batch axis 233 for operand 350 fallback batch axis 233 for operand 365 fallback batch axis 233 for operand 387 fallback batch axis 233 for operand 398 fallback batch axis 233 for operand 413 fallback batch axis 233 for operand 435 fallback batch axis 233 for operand 446 fallback batch axis 233 for operand 461 fallback batch axis 233 for operand 483 fallback batch axis 233 for operand 494 fallback batch axis 233 for operand 509 fallback batch axis 233 for operand 531 fallback batch axis 233 for operand 542

model | 模型 | モデル

original model

how to reproduce | 复现步骤 | 再現方法

1../pnnx gfpganv1_clean_model.pt inputshape=[1,3,512,512] fp16=0 2.pnnx版本20240226 3.

请问类似这样的报错，我怎么定位到pytorch的具体op上？196，207这些在ncnn的param里和pt里都搜不到

Baiyuetribe commented 2 months ago

所以这个是如何定位的？我看你已经出来gfpgan1.4的demo了

zhang0557kui commented 2 months ago

同问

XdpAreKid commented 2 months ago

https://github.com/magicse/GFPGANv1.3-to-ncnn 用的这个仓库里的代码，然后需要把里面一个层改了，具体哪个层有点忘了

zhangkui @.***> 于2024年9月1日周日 00:12写道：

同问

— Reply to this email directly, view it on GitHub https://github.com/Tencent/ncnn/issues/5462#issuecomment-2322949198, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFK2FPCHDYKBNCMDFEZ7MNTZUHTPHAVCNFSM6AAAAABHYCI4T6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMRSHE2DSMJZHA . You are receiving this because you authored the thread.Message ID: @.***>

Tencent / ncnn