chunked/content_aware_chunking/_custom_models/chunk_12.txt

Using another output format is fine as long as you are planning on using your own
training loop or another library for training.
Now that we have our model class, let's create one:
py
resnet50d = ResnetModelForImageClassification(resnet50d_config)
Again, you can use any of the methods of [PreTrainedModel], like [~PreTrainedModel.save_pretrained] or
[~PreTrainedModel.push_to_hub]. We will use the second in the next section, and see how to push the model weights
with the code of our model.