tanganke's picture
Upload folder using huggingface_hub
e5d741e verified
metadata
base_model:
  - openai/clip-vit-large-patch14
datasets:
  - clip-benchmark/wds_fer2013
metrics:
  - accuracy

Model Card

Training Details

Adam Optimizer with a constant learning rate 5e-6 for 4000 steps training (batch_size=128). Only the vision encoder is fine-tuned.

Evaluation Results

Test set accuracy:

  • pre-trained: 0.49958205223083496
  • fine-tuned: 0.7592644095420837