Error when loading pretrained coatnet

[Soongja]

Hi, I’d like to report another issue.

When I run the following code:

model = timm_3d.create_model("coatnet_0_rw_224.sw_in1k", pretrained=True, in_chans=1, num_classes=14)

I get the following error.

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/sj/lib/python3.10/site-packages/timm_3d/models/_factory.py", line 117, in create_model
    model = create_fn(
  File "/sj/lib/python3.10/site-packages/timm_3d/models/maxxvit.py", line 2116, in coatnet_0_rw_224
    return _create_maxxvit('coatnet_0_rw_224', pretrained=pretrained, **kwargs)
  File "/sj/lib/python3.10/site-packages/timm_3d/models/maxxvit.py", line 1858, in _create_maxxvit
    return build_model_with_cfg(
  File "/sj/lib/python3.10/site-packages/timm_3d/models/_builder.py", line 439, in build_model_with_cfg
    load_pretrained(
  File "/sj/lib/python3.10/site-packages/timm_3d/models/_builder.py", line 234, in load_pretrained
    state_dict[weight_name] = adapt_input_conv(in_chans, state_dict[weight_name])
  File "/sj/lib/python3.10/site-packages/timm_3d/models/_manipulate.py", line 259, in adapt_input_conv
    O, I, J, K = conv_weight.shape
ValueError: too many values to unpack (expected 4)

The problem does not occur when I set pretrained=False.

Could you please check if it can be fixed?