Bird Species Classifier (ViT)

Fine-tuned Vision Transformer (ViT-base-patch16-224) on the CUB-200-2011 dataset for 200-class bird species classification.

Model Details

  • Architecture: ViT-base-patch16-224 (Google)
  • Dataset: CUB-200-2011 (11,788 images, 200 species)
  • Training: Transfer learning with AdamW optimizer, 5 epochs
  • Test Accuracy: 99.93%
  • Framework: PyTorch + HuggingFace Transformers

Author

Mohamed Fiyaz

Downloads last month
33
Safetensors
Model size
86M params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for MohamedFiyaz/bird-species-classifier

Finetuned
(2107)
this model

Space using MohamedFiyaz/bird-species-classifier 1