README.md

---
base_model:
- openai/clip-vit-base-patch32
datasets:
- clip-benchmark/wds_fer2013
metrics:
- accuracy
---

# Model Card

## Training Details

Adam Optimizer with a constant learning rate 1e-5 for 4000 steps training (batch_size=128).
Only the vision encoder is fine-tuned.

![alt text](image.png)

## Evaluation Results

Test set accuracy:

- pre-trained: 0.4125104546546936
- fine-tuned: 0.7160769104957581