Pre-trained and fine-tuned models for computer vision, natural language processing, and medical AI.
DINOv2 ViT-Large finetuned on CT-RATE for chest CT feature extraction with anatomically-aware cropping.
Vision Transformer for neuropathology, pretrained using DinoMX (DINO + iBOT). Linear probe accuracy: 80.17%, KNN accuracy: 83.76%.
Lightweight medical LLM. 13.76% improvement over base TinyLlama across 3 medical benchmarks.
Medical LLM achieving 68.2% across USMLE/AIIMS/NEET. 4.42% improvement over base Mixtral.
Medical LLM based on LLaMA 2 7B, fine-tuned on medical datasets.