3D Pose Size mismatch

[Gnoot01]

Hi, I am getting the following error even after following the instructions here. I followed to also use the Halpe 26 model to generate the Halpe 26 json file as input to the 3D MotionBERT.

Should i be using config file MB_ft_h36m_global.yaml instead of MB_ft_h36m_global_lite.yaml (instructions)?

File "/mnt/c/Users/MotionBERT-main/.wsl_venv/lib/python3.10/site-packages/torch/nn/modules/module.py", line 2215, in load_state_dict
    raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for DataParallel:
        size mismatch for module.temp_embed: copying a param with shape torch.Size([1, 243, 1, 512]) from checkpoint, the shape in current model is torch.Size([1, 243, 1, 256]).
        size mismatch for module.pos_embed: copying a param with shape torch.Size([1, 17, 512]) from checkpoint, the shape in current model is torch.Size([1, 17, 256]).
        size mismatch for module.joints_embed.weight: copying a param with shape torch.Size([512, 3]) from checkpoint, the shape in current model is torch.Size([256, 3]).
        size mismatch for module.joints_embed.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.0.norm1_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.0.norm1_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.0.norm1_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.0.norm1_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.0.attn_s.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_st.0.attn_s.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.0.attn_s.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_st.0.attn_s.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_st.0.attn_t.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_st.0.attn_t.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.0.attn_t.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_st.0.attn_t.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_st.0.norm2_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.0.norm2_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.0.norm2_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.0.norm2_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.0.mlp_s.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_st.0.mlp_s.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_st.0.mlp_s.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.0.mlp_t.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_st.0.mlp_t.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_st.0.mlp_t.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.1.norm1_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.1.norm1_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.1.norm1_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.1.norm1_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.1.attn_s.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_st.1.attn_s.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.1.attn_s.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_st.1.attn_s.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_st.1.attn_t.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_st.1.attn_t.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.1.attn_t.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_st.1.attn_t.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_st.1.norm2_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.1.norm2_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.1.norm2_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.1.norm2_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.1.mlp_s.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_st.1.mlp_s.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_st.1.mlp_s.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.1.mlp_t.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_st.1.mlp_t.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_st.1.mlp_t.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.2.norm1_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.2.norm1_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.2.norm1_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.2.norm1_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.2.attn_s.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_st.2.attn_s.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.2.attn_s.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_st.2.attn_s.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_st.2.attn_t.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_st.2.attn_t.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.2.attn_t.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_st.2.attn_t.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_st.2.norm2_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.2.norm2_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.2.norm2_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.2.norm2_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.2.mlp_s.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_st.2.mlp_s.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_st.2.mlp_s.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.2.mlp_t.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_st.2.mlp_t.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_st.2.mlp_t.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.3.norm1_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.3.norm1_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.3.norm1_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.3.norm1_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.3.attn_s.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_st.3.attn_s.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.3.attn_s.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_st.3.attn_s.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_st.3.attn_t.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_st.3.attn_t.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.3.attn_t.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_st.3.attn_t.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_st.3.norm2_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.3.norm2_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.3.norm2_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.3.norm2_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.3.mlp_s.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_st.3.mlp_s.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_st.3.mlp_s.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.3.mlp_t.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_st.3.mlp_t.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_st.3.mlp_t.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.4.norm1_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.4.norm1_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.4.norm1_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.4.norm1_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.4.attn_s.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_st.4.attn_s.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.4.attn_s.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_st.4.attn_s.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_st.4.attn_t.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_st.4.attn_t.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.4.attn_t.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_st.4.attn_t.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_st.4.norm2_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.4.norm2_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.4.norm2_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.4.norm2_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.4.mlp_s.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_st.4.mlp_s.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_st.4.mlp_s.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_st.4.mlp_t.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_st.4.mlp_t.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_st.4.mlp_t.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.0.norm1_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.0.norm1_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.0.norm1_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.0.norm1_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.0.attn_s.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_ts.0.attn_s.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.0.attn_s.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_ts.0.attn_s.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_ts.0.attn_t.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_ts.0.attn_t.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.0.attn_t.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_ts.0.attn_t.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_ts.0.norm2_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.0.norm2_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.0.norm2_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.0.norm2_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.0.mlp_s.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_ts.0.mlp_s.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_ts.0.mlp_s.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.0.mlp_t.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_ts.0.mlp_t.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_ts.0.mlp_t.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.1.norm1_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.1.norm1_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.1.norm1_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.1.norm1_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.1.attn_s.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_ts.1.attn_s.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.1.attn_s.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_ts.1.attn_s.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_ts.1.attn_t.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_ts.1.attn_t.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.1.attn_t.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_ts.1.attn_t.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_ts.1.norm2_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.1.norm2_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.1.norm2_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.1.norm2_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.1.mlp_s.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_ts.1.mlp_s.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_ts.1.mlp_s.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.1.mlp_t.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_ts.1.mlp_t.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_ts.1.mlp_t.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.2.norm1_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.2.norm1_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.2.norm1_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.2.norm1_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.2.attn_s.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_ts.2.attn_s.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.2.attn_s.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_ts.2.attn_s.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_ts.2.attn_t.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_ts.2.attn_t.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.2.attn_t.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_ts.2.attn_t.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_ts.2.norm2_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.2.norm2_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.2.norm2_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.2.norm2_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.2.mlp_s.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_ts.2.mlp_s.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_ts.2.mlp_s.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.2.mlp_t.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_ts.2.mlp_t.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_ts.2.mlp_t.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.3.norm1_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.3.norm1_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.3.norm1_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.3.norm1_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.3.attn_s.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_ts.3.attn_s.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.3.attn_s.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_ts.3.attn_s.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_ts.3.attn_t.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_ts.3.attn_t.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.3.attn_t.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_ts.3.attn_t.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_ts.3.norm2_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.3.norm2_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.3.norm2_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.3.norm2_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.3.mlp_s.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_ts.3.mlp_s.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_ts.3.mlp_s.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.3.mlp_t.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_ts.3.mlp_t.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_ts.3.mlp_t.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.4.norm1_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.4.norm1_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.4.norm1_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.4.norm1_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.4.attn_s.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_ts.4.attn_s.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.4.attn_s.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_ts.4.attn_s.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_ts.4.attn_t.proj.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([256, 256]).
        size mismatch for module.blocks_ts.4.attn_t.proj.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.4.attn_t.qkv.weight: copying a param with shape torch.Size([1536, 512]) from checkpoint, the shape in current model is torch.Size([768, 256]).
        size mismatch for module.blocks_ts.4.attn_t.qkv.bias: copying a param with shape torch.Size([1536]) from checkpoint, the shape in current model is torch.Size([768]).
        size mismatch for module.blocks_ts.4.norm2_s.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.4.norm2_s.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.4.norm2_t.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.4.norm2_t.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.4.mlp_s.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_ts.4.mlp_s.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_ts.4.mlp_s.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.blocks_ts.4.mlp_t.fc1.weight: copying a param with shape torch.Size([1024, 512]) from checkpoint, the shape in current model is torch.Size([1024, 256]).
        size mismatch for module.blocks_ts.4.mlp_t.fc2.weight: copying a param with shape torch.Size([512, 1024]) from checkpoint, the shape in current model is torch.Size([256, 1024]).
        size mismatch for module.blocks_ts.4.mlp_t.fc2.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.norm.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.norm.bias: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([256]).
        size mismatch for module.pre_logits.fc.weight: copying a param with shape torch.Size([512, 512]) from checkpoint, the shape in current model is torch.Size([512, 256]).
        size mismatch for module.ts_attn.0.weight: copying a param with shape torch.Size([2, 1024]) from checkpoint, the shape in current model is torch.Size([2, 512]).
        size mismatch for module.ts_attn.1.weight: copying a param with shape torch.Size([2, 1024]) from checkpoint, the shape in current model is torch.Size([2, 512]).
        size mismatch for module.ts_attn.2.weight: copying a param with shape torch.Size([2, 1024]) from checkpoint, the shape in current model is torch.Size([2, 512]).
        size mismatch for module.ts_attn.3.weight: copying a param with shape torch.Size([2, 1024]) from checkpoint, the shape in current model is torch.Size([2, 512]).
        size mismatch for module.ts_attn.4.weight: copying a param with shape torch.Size([2, 1024]) from checkpoint, the shape in current model is torch.Size([2, 512]).