ResNet-50 ImageNet-20 Submission Demo
This is a Hugging Face-compatible image-classification submission.
Labels
apple, orange, bee, butterfly, camel, elephant, lion, tiger, squirrel, shark, turtle, crab, lobster, bicycle, pickup_truck, tractor, train, keyboard, television, wardrobe
Original ImageNet class indices used for filtering
{'apple': 948, 'orange': 950, 'bee': 309, 'butterfly': 323, 'camel': 354, 'elephant': 386, 'lion': 291, 'tiger': 292, 'squirrel': 335, 'shark': 2, 'turtle': 33, 'crab': 118, 'lobster': 122, 'bicycle': 444, 'pickup_truck': 717, 'tractor': 866, 'train': 820, 'keyboard': 508, 'television': 851, 'wardrobe': 893}
Required loading interface
1from transformers import AutoImageProcessor, AutoModelForImageClassification
2
3processor = AutoImageProcessor.from_pretrained(repo_id, revision=commit_hash)
4model = AutoModelForImageClassification.from_pretrained(
5 repo_id,
6 revision=commit_hash,
7 trust_remote_code=False,
8)
