[BUG] Error when downloading dinov2_* zoo models

[swheaton]

System information

• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Mac OSX
• Python version (python --version): 3.8.10
• FiftyOne version (fiftyone --version): 0.22.1
• FiftyOne installed from (pip or source): pip

Describe the problem

I get an error ValueError: Layer 'head' is an Identity layer. Use previous layer. when downloading DINOv2 zoo models.. Below is one, I tried most of the others and got the same error.

>>> m=fo.zoo.load_zoo_model('dinov2-vits14-torch')
Downloading: "https://dl.fbaipublicfiles.com/dinov2/dinov2_vits14/dinov2_vits14_pretrain.pth" to /Users/stuart/.cache/torch/hub/checkpoints/dinov2_vits14_pretrain.pth
100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 84.2M/84.2M [00:01<00:00, 49.3MB/s]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/stuart/dev/fiftyone_venv2/lib/python3.8/site-packages/fiftyone/zoo/models/__init__.py", line 253, in load_zoo_model
    model = fom.load_model(config_dict, model_path=model_path, **kwargs)
  File "/Users/stuart/dev/fiftyone_venv2/lib/python3.8/site-packages/fiftyone/core/models.py", line 1777, in load_model
    return config.build()
  File "/Users/stuart/dev/fiftyone_venv2/lib/python3.8/site-packages/eta/core/learning.py", line 296, in build
    return self._model_cls(self.config)
  File "/Users/stuart/dev/fiftyone_venv2/lib/python3.8/site-packages/fiftyone/utils/torch.py", line 549, in __init__
    TorchEmbeddingsMixin.__init__(
  File "/Users/stuart/dev/fiftyone_venv2/lib/python3.8/site-packages/fiftyone/utils/torch.py", line 271, in __init__
    embeddings_layer = SaveLayerTensor(model, layer_name)
  File "/Users/stuart/dev/fiftyone_venv2/lib/python3.8/site-packages/fiftyone/utils/torch.py", line 1010, in __init__
    self._setup(model, layer_name)
  File "/Users/stuart/dev/fiftyone_venv2/lib/python3.8/site-packages/fiftyone/utils/torch.py", line 1035, in _setup
    raise ValueError(
ValueError: Layer 'head' is an Identity layer. Use previous layer.

Code to reproduce issue

import fiftyone as fo
model=fo.zoo.load_zoo_model('dinov2-vits14-torch')

What areas of FiftyOne does this bug affect?

• App: FiftyOne application issue
• Core: Core Python library issue
• Server: FiftyOne server issue
• Zoo: FiftyOne Zoo

Willingness to contribute

The FiftyOne Community encourages bug fix contributions. Would you or another member of your organization be willing to contribute a fix for this bug to the FiftyOne codebase?

• Yes. I can contribute a fix for this bug independently
• Yes. I would be willing to contribute a fix for this bug with guidance from the FiftyOne community
• No. I cannot contribute a bug fix at this time

[brimoor]

This was resolved by #3690 which was released in fiftyone==0.22.2. Unsure why GH did not automatically close this.