TRANSUNET SAYS KEY ERROR

[abhinav70291]

I am trying to use the transunet for as the --net argument. However it throws an error as depited in the image.The error reads as :"Traceback (most recent call last): File "/content/segtran/code/train2d.py", line 976, in transunet_config = TransUNet_CONFIGS[args.backbone_type] KeyError: 'eff-b4'".

Even if I change the backbone to eff-b1 or resnet101, a new key error is raised for those backbones.Please help me out

[abhinav70291]

So I changed the argument to Vit b 16 and now it says :Key error :n_skip. Please help me out @askerlee @nqn @segtran

[askerlee]

The supported backbones are: CONFIGS = { 'ViT-B_16': configs.get_b16_config(), 'ViT-B_32': configs.get_b32_config(), 'ViT-L_16': configs.get_l16_config(), 'ViT-L_32': configs.get_l32_config(), 'ViT-H_14': configs.get_h14_config(), 'R50-ViT-B_16': configs.get_r50_b16_config(), 'R50-ViT-L_16': configs.get_r50_l16_config(), 'R50-ViT-L_32': configs.get_r50_l32_config(), 'testing': configs.get_testing(), } (This dict is in vit_seg_modeling.py.)

Since you've found the correct backbone_type, the second error looks like a different version of ml_collections? Maybe it's too new. Yes it's my bad that I didn't specify the version number of packages in requirements.txt.