search

arcadelab / FastSAM3D

Code for "FastSAM3D: An Efficient Segment Anything Model for 3D Volumetric Medical Images"
https://arxiv.org/abs/2403.09827
Apache License 2.0
68 stars 4 forks source link

state dict error #1

Closed Overflowu7 closed 7 months ago

Overflowu7 commented 8 months ago

When i run preparelable.sh and use sam_med3d_turbo.pth with pretrain, the problem of parameter mismatch arises.

Have you ever had such a problem before?

/FastSAM3D-main/preparelabel.py get 1 datasets Dataset size: 20 Dataset size: 20 Traceback (most recent call last): File "/home/FastSAM3D-main/preparelabel.py", line 82, in sam_model_tune.load_state_dict(state_dict) File "/home/w/.conda/envs/SAM/lib/python3.10/site-packages/torch/nn/modules/module.py", line 2153, in load_state_dict raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format( RuntimeError: Error(s) in loading state_dict for Sam3D: Unexpected key(s) in state_dict: "image_encoder.blocks.0.norm1.weight", "image_encoder.blocks.0.norm1.bias", "image_encoder.blocks.0.attn.rel_pos_d", "image_encoder.blocks.0.attn.rel_pos_h", "image_encoder.blocks.0.attn.rel_pos_w", "image_encoder.blocks.0.attn.qkv.weight", "image_encoder.blocks.0.attn.qkv.bias", "image_encoder.blocks.0.attn.proj.weight", "image_encoder.blocks.0.attn.proj.bias", "image_encoder.blocks.1.norm1.weight", "image_encoder.blocks.1.norm1.bias", "image_encoder.blocks.1.attn.rel_pos_d", "image_encoder.blocks.1.attn.rel_pos_h", "image_encoder.blocks.1.attn.rel_pos_w", "image_encoder.blocks.1.attn.qkv.weight", "image_encoder.blocks.1.attn.qkv.bias", "image_encoder.blocks.1.attn.proj.weight", "image_encoder.blocks.1.attn.proj.bias", "image_encoder.blocks.2.norm1.weight", "image_encoder.blocks.2.norm1.bias", "image_encoder.blocks.2.attn.rel_pos_d", "image_encoder.blocks.2.attn.rel_pos_h", "image_encoder.blocks.2.attn.rel_pos_w", "image_encoder.blocks.2.attn.qkv.weight", "image_encoder.blocks.2.attn.qkv.bias", "image_encoder.blocks.2.attn.proj.weight", "image_encoder.blocks.2.attn.proj.bias", "image_encoder.blocks.3.norm1.weight", "image_encoder.blocks.3.norm1.bias", "image_encoder.blocks.3.attn.rel_pos_d", "image_encoder.blocks.3.attn.rel_pos_h", "image_encoder.blocks.3.attn.rel_pos_w", "image_encoder.blocks.3.attn.qkv.weight", "image_encoder.blocks.3.attn.qkv.bias", "image_encoder.blocks.3.attn.proj.weight", "image_encoder.blocks.3.attn.proj.bias", "image_encoder.blocks.4.norm1.weight", "image_encoder.blocks.4.norm1.bias", "image_encoder.blocks.4.attn.rel_pos_d", "image_encoder.blocks.4.attn.rel_pos_h", "image_encoder.blocks.4.attn.rel_pos_w", "image_encoder.blocks.4.attn.qkv.weight", "image_encoder.blocks.4.attn.qkv.bias", "image_encoder.blocks.4.attn.proj.weight", "image_encoder.blocks.4.attn.proj.bias", "image_encoder.blocks.5.norm1.weight", "image_encoder.blocks.5.norm1.bias", "image_encoder.blocks.5.attn.rel_pos_d", "image_encoder.blocks.5.attn.rel_pos_h", "image_encoder.blocks.5.attn.rel_pos_w", "image_encoder.blocks.5.attn.qkv.weight", "image_encoder.blocks.5.attn.qkv.bias", "image_encoder.blocks.5.attn.proj.weight", "image_encoder.blocks.5.attn.proj.bias".

skill-diver commented 8 months ago

We just fixed, thank you for your feedback.

chuyhu commented 8 months ago

We just fixed, thank you for your feedback.

Hello, it seems that this issue has not been resolved.

swedfr commented 8 months ago

I have do some more fix on that, currently the preparelabel.sh should be correct. Thank you!