checkpoint is not compatible with the network architecture
ZhengzxDev opened this issue · comments
your model is perfectly satisfied with our requirements, but after several attempts, it seems that the checkpoint file that you provided is not compatible with the code. can you provide the latest pretrain-model?
me too. how can i get pre-trained model?
I am getting a similar error:
normal init for last conv
Trying to load from: ./ckpts/model_btmup/
Traceback (most recent call last):
File "calculate_mupots_btmup.py", line 30, in
M.Saver(model).restore('./ckpts/model_btmup/')
File "/home/jpsml/anaconda3/envs/3dmpp/lib/python3.6/site-packages/TorchSUL/Model.py", line 156, in restore
self.model.load_state_dict(state_dict, strict=strict)
File "/home/jpsml/anaconda3/envs/3dmpp/lib/python3.6/site-packages/torch/nn/modules/module.py", line 1037, in load_state_dict
load(self)
File "/home/jpsml/anaconda3/envs/3dmpp/lib/python3.6/site-packages/torch/nn/modules/module.py", line 1035, in load
load(child, prefix + name + '.')
File "/home/jpsml/anaconda3/envs/3dmpp/lib/python3.6/site-packages/torch/nn/modules/module.py", line 1035, in load
load(child, prefix + name + '.')
File "/home/jpsml/anaconda3/envs/3dmpp/lib/python3.6/site-packages/torch/nn/modules/module.py", line 1035, in load
load(child, prefix + name + '.')
[Previous line repeated 1 more time]
File "/home/jpsml/anaconda3/envs/3dmpp/lib/python3.6/site-packages/torch/nn/modules/module.py", line 1032, in load
state_dict, prefix, local_metadata, True, missing_keys, unexpected_keys, error_msgs)
File "/home/jpsml/anaconda3/envs/3dmpp/lib/python3.6/site-packages/TorchSUL/Layers.py", line 106, in _load_from_state_dict
raise Exception('No weight found for layer:', prefix)
Exception: ('No weight found for layer:', 'backbone.lastfuse.c_all.conv.')
Please try to downgrade your TorchSUL to 0.1.28