search

xinntao / EDVR

Winning Solution in NTIRE19 Challenges on Video Restoration and Enhancement (CVPR19 Workshops) - Video Restoration with Enhanced Deformable Convolutional Networks. EDVR has been merged into BasicSR and this repo is a mirror of BasicSR.
https://github.com/xinntao/BasicSR
1.48k stars 320 forks source link

Missing key(s) in state_dict: "tsa_fusion.tAtt_1.weight" #71

Closed robotzheng closed 5 years ago

robotzheng commented 5 years ago

train a model with tsa=true from a pretrained model with tsa=false .

RuntimeError: Error(s) in loading state_dict for EDVR: Missing key(s) in state_dict: "tsa_fusion.tAtt_1.weight", "tsa_fusion.tAtt_1.bias", "tsa_fusion.tAtt_2.weight", "tsa_fusion.tAtt_2.bias", "tsa_fusion.fea_fusion.weight", "tsa_fusion.fea_fusion.bias", "tsa_fusion.sAtt_1.weight", "tsa_fusion.sAtt_1.bias", "tsa_fusion.sAtt_2.weight", "tsa_fusion.sAtt_2.bias", "tsa_fusion.sAtt_3.weight", "tsa_fusion.sAtt_3.bias", "tsa_fusion.sAtt_4.weight", "tsa_fusion.sAtt_4.bias", "tsa_fusion.sAtt_5.weight", "tsa_fusion.sAtt_5.bias", "tsa_fusion.sAtt_L1.weight", "tsa_fusion.sAtt_L1.bias", "tsa_fusion.sAtt_L2.weight", "tsa_fusion.sAtt_L2.bias", "tsa_fusion.sAtt_L3.weight", "tsa_fusion.sAtt_L3.bias", "tsa_fusion.sAtt_add_1.weight", "tsa_fusion.sAtt_add_1.bias", "tsa_fusion.sAtt_add_2.weight", "tsa_fusion.sAtt_add_2.bias". Unexpected key(s) in state_dict: "tsa_fusion.weight", "tsa_fusion.bias".

xinntao commented 5 years ago

A model with tsa=true has more weights. You can set strict_load: false, which only loads the existing weights.

robotzheng commented 5 years ago

thanks.