Robust Fine-tuning of Zero-shot Models via Variance Reduction

VLM
OODD
Re-assign community