Unable to load model with SWA. Unexpected key in state_dict

[codezbq]

Traceback (most recent calls WITHOUT Sacred internals):
  File "/home/bingqing.zhao/PaSST/ex_fsd50k.py", line 480, in default_command
    return main()
  File "/home/bingqing.zhao/PaSST/ex_fsd50k.py", line 303, in main
    modul = M.load_from_checkpoint(checkpoint_path=checkpoint_path,experiment=ex)
  File "/home/bingqing.zhao/PaSST/src/pytorch-lightning/pytorch_lightning/core/saving.py", line 157, in load_from_checkpoint
    model = cls._load_model_state(checkpoint, strict=strict, **kwargs)
  File "/home/bingqing.zhao/PaSST/src/pytorch-lightning/pytorch_lightning/core/saving.py", line 205, in _load_model_state
    model.load_state_dict(checkpoint['state_dict'], strict=strict)
  File "/home/bingqing.zhao/.conda/envs/ba3l/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1498, in load_state_dict
    self.__class__.__name__, "\n\t".join(error_msgs)))
RuntimeError: Error(s) in loading state_dict for M:
	Unexpected key(s) in state_dict: "net_swa.cls_token", "net_swa.dist_token", "net_swa.new_pos_embed", "net_swa.freq_new_pos_embed", "net_swa.time_new_pos_embed", "net_swa.patch_embed.proj.weight", "net_swa.patch_embed.proj.bias", "net_swa.blocks.0.norm1.weight", "net_swa.blocks.0.norm1.bias", "net_swa.blocks.0.attn.qkv.weight", "net_swa.blocks.0.attn.qkv.bias", "net_swa.blocks.0.attn.proj.weight", "net_swa.blocks.0.attn.proj.bias", "net_swa.blocks.0.norm2.weight", "net_swa.blocks.0.norm2.bias", "net_swa.blocks.0.mlp.fc1.weight", "net_swa.blocks.0.mlp.fc1.bias", "net_swa.blocks.0.mlp.fc2.weight", "net_swa.blocks.0.mlp.fc2.bias", "net_swa.blocks.1.norm1.weight", "net_swa.blocks.1.norm1.bias", "net_swa.blocks.1.attn.qkv.weight", "net_swa.blocks.1.attn.qkv.bias", "net_swa.blocks.1.attn.proj.weight", "net_swa.blocks.1.attn.proj.bias", "net_swa.blocks.1.norm2.weight", "net_swa.blocks.1.norm2.bias", "net_swa.blocks.1.mlp.fc1.weight", "net_swa.blocks.1.mlp.fc1.bias", "net_swa.blocks.1.mlp.fc2.weight", "net_swa.blocks.1.mlp.fc2.bias", "net_swa.blocks.2.norm1.weight", "net_swa.blocks.2.norm1.bias", "net_swa.blocks.2.attn.qkv.weight", "net_swa.blocks.2.attn.qkv.bias", "net_swa.blocks.2.attn.proj.weight", "net_swa.blocks.2.attn.proj.bias", "net_swa.blocks.2.norm2.weight", "net_swa.blocks.2.norm2.bias", "net_swa.blocks.2.mlp.fc1.weight", "net_swa.blocks.2.mlp.fc1.bias", "net_swa.blocks.2.mlp.fc2.weight", "net_swa.blocks.2.mlp.fc2.bias", "net_swa.blocks.3.norm1.weight", "net_swa.blocks.3.norm1.bias", "net_swa.blocks.3.attn.qkv.weight", "net_swa.blocks.3.attn.qkv.bias", "net_swa.blocks.3.attn.proj.weight", "net_swa.blocks.3.attn.proj.bias", "net_swa.blocks.3.norm2.weight", "net_swa.blocks.3.norm2.bias", "net_swa.blocks.3.mlp.fc1.weight", "net_swa.blocks.3.mlp.fc1.bias", "net_swa.blocks.3.mlp.fc2.weight", "net_swa.blocks.3.mlp.fc2.bias", "net_swa.blocks.4.norm1.weight", "net_swa.blocks.4.norm1.bias", "net_swa.blocks.4.attn.qkv.weight", "net_swa.blocks.4.attn.qkv.bias", "net_swa.blocks.4.attn.proj.weight", "net_swa.blocks.4.attn.proj.bias", "net_swa.blocks.4.norm2.weight", "net_swa.blocks.4.norm2.bias", "net_swa.blocks.4.mlp.fc1.weight", "net_swa.blocks.4.mlp.fc1.bias", "net_swa.blocks.4.mlp.fc2.weight", "net_swa.blocks.4.mlp.fc2.bias", "net_swa.blocks.5.norm1.weight", "net_swa.blocks.5.norm1.bias", "net_swa.blocks.5.attn.qkv.weight", "net_swa.blocks.5.attn.qkv.bias", "net_swa.blocks.5.attn.proj.weight", "net_swa.blocks.5.attn.proj.bias", "net_swa.blocks.5.norm2.weight", "net_swa.blocks.5.norm2.bias", "net_swa.blocks.5.mlp.fc1.weight", "net_swa.blocks.5.mlp.fc1.bias", "net_swa.blocks.5.mlp.fc2.weight", "net_swa.blocks.5.mlp.fc2.bias", "net_swa.blocks.6.norm1.weight", "net_swa.blocks.6.norm1.bias", "net_swa.blocks.6.attn.qkv.weight", "net_swa.blocks.6.attn.qkv.bias", "net_swa.blocks.6.attn.proj.weight", "net_swa.blocks.6.attn.proj.bias", "net_swa.blocks.6.norm2.weight", "net_swa.blocks.6.norm2.bias", "net_swa.blocks.6.mlp.fc1.weight", "net_swa.blocks.6.mlp.fc1.bias", "net_swa.blocks.6.mlp.fc2.weight", "net_swa.blocks.6.mlp.fc2.bias", "net_swa.blocks.7.norm1.weight", "net_swa.blocks.7.norm1.bias", "net_swa.blocks.7.attn.qkv.weight", "net_swa.blocks.7.attn.qkv.bias", "net_swa.blocks.7.attn.proj.weight", "net_swa.blocks.7.attn.proj.bias", "net_swa.blocks.7.norm2.weight", "net_swa.blocks.7.norm2.bias", "net_swa.blocks.7.mlp.fc1.weight", "net_swa.blocks.7.mlp.fc1.bias", "net_swa.blocks.7.mlp.fc2.weight", "net_swa.blocks.7.mlp.fc2.bias", "net_swa.blocks.8.norm1.weight", "net_swa.blocks.8.norm1.bias", "net_swa.blocks.8.attn.qkv.weight", "net_swa.blocks.8.attn.qkv.bias", "net_swa.blocks.8.attn.proj.weight", "net_swa.blocks.8.attn.proj.bias", "net_swa.blocks.8.norm2.weight", "net_swa.blocks.8.norm2.bias", "net_swa.blocks.8.mlp.fc1.weight", "net_swa.blocks.8.mlp.fc1.bias", "net_swa.blocks.8.mlp.fc2.weight", "net_swa.blocks.8.mlp.fc2.bias", "net_swa.blocks.9.norm1.weight", "net_swa.blocks.9.norm1.bias", "net_swa.blocks.9.attn.qkv.weight", "net_swa.blocks.9.attn.qkv.bias", "net_swa.blocks.9.attn.proj.weight", "net_swa.blocks.9.attn.proj.bias", "net_swa.blocks.9.norm2.weight", "net_swa.blocks.9.norm2.bias", "net_swa.blocks.9.mlp.fc1.weight", "net_swa.blocks.9.mlp.fc1.bias", "net_swa.blocks.9.mlp.fc2.weight", "net_swa.blocks.9.mlp.fc2.bias", "net_swa.blocks.10.norm1.weight", "net_swa.blocks.10.norm1.bias", "net_swa.blocks.10.attn.qkv.weight", "net_swa.blocks.10.attn.qkv.bias", "net_swa.blocks.10.attn.proj.weight", "net_swa.blocks.10.attn.proj.bias", "net_swa.blocks.10.norm2.weight", "net_swa.blocks.10.norm2.bias", "net_swa.blocks.10.mlp.fc1.weight", "net_swa.blocks.10.mlp.fc1.bias", "net_swa.blocks.10.mlp.fc2.weight", "net_swa.blocks.10.mlp.fc2.bias", "net_swa.blocks.11.norm1.weight", "net_swa.blocks.11.norm1.bias", "net_swa.blocks.11.attn.qkv.weight", "net_swa.blocks.11.attn.qkv.bias", "net_swa.blocks.11.attn.proj.weight", "net_swa.blocks.11.attn.proj.bias", "net_swa.blocks.11.norm2.weight", "net_swa.blocks.11.norm2.bias", "net_swa.blocks.11.mlp.fc1.weight", "net_swa.blocks.11.mlp.fc1.bias", "net_swa.blocks.11.mlp.fc2.weight", "net_swa.blocks.11.mlp.fc2.bias", "net_swa.norm.weight", "net_swa.norm.bias", "net_swa.head.0.weight", "net_swa.head.0.bias", "net_swa.head.1.weight", "net_swa.head.1.bias", "net_swa.head_dist.weight", "net_swa.head_dist.bias". 

It seems like every part of the network was append the name "swa", and it confuse me a lot that how to resume training when I use the SWA. Please help