[BUG] tf_efficientnetv2_xl_in21k doesnt load

[emoen]

Describe the bug unable to load efficientNetV2 with xl weights

To Reproduce Steps to reproduce the behavior:

timm.create_model('tf_efficientnetv2_xl_in21k', pretrained=True, in_chans=3, num_classes=1) fails to load the model with error:

---

RuntimeError Traceback (most recent call last)

in ----> 1 timm.create_model('tf_efficientnetv2_xl_in21k', pretrained=True, in_chans=3, num_classes=1000) ~/.local/lib/python3.6/site-packages/timm/models/factory.py in create_model(model_name, pretrained, checkpoint_path, scriptable, exportable, no_jit, **kwargs) 76 create_fn = model_entrypoint(model_name) 77 else: ---> 78 raise RuntimeError('Unknown model (%s)' % model_name) 79 80 with set_layer_config(scriptable=scriptable, exportable=exportable, no_jit=no_jit): RuntimeError: Unknown model (tf_efficientnetv2_xl_in21k) **Expected behavior** loading large: >timm.create_model('tf_efficientnetv2_l_in21k', pretrained=True, in_chans=3, num_classes=1000) EfficientNet( (conv_stem): Conv2dSame(3, 32, kernel_size=(3, 3), stride=(2, 2), bias=False) (bn1): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True) (act1): SiLU(inplace=True) ... (classifier): Linear(in_features=1280, out_features=1000, bias=True) ) **Desktop (please complete the following information):** - OS: centO, running jupyter

[rwightman]

@emoen check your version of timm, it works fine with curent master branch and pypi release. You have an old one in your environment

[emoen]

version I had: timm 0.4.12, new version: timm-0.5.4

[ChristopherMarais]

I am getting the same issue with maxvit_rmlp_base_rw_224.sw_in12k from here. I have also checked and I am using timm version 0.6.12

[hhhheep]

I am getting the same issue, and solved it with pip install timm --upgrade.