vit-base-oxford-iiit-pets

This model is a fine-tuned version of google/vit-base-patch16-224 on the pcuenq/oxford-pets dataset. It achieves the following results on the evaluation set:

• Loss: 0.1893
• Accuracy: 0.9405

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0003
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.3976	1.0	370	0.2921	0.9364
0.2273	2.0	740	0.2257	0.9445
0.1742	3.0	1110	0.2102	0.9445
0.1352	4.0	1480	0.2023	0.9459
0.1326	5.0	1850	0.2006	0.9459

Framework versions

• Transformers 4.50.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1

Zero-Shot classification model

This section compares the performance of a zero-shot model (openai/clip-vit-large-patch14) on the Oxford Pets dataset (pcuenq/oxford-pets).

• Model used: openai/clip-vit-large-patch14
• Dataset: pcuenq/oxford-pets (train split)
• Evaluation Task: Zero-Shot Image Classification
• Candidate Labels: 37 pet breeds from the dataset

Results:

Zero-Shot Evaluation mit CLIP:

• Accuracy: 0.8800
• Precision: 0.8768
• Recall: 0.8800

Evaluated using Hugging Face transformers pipeline and sklearn.metrics on the full training set.