application file and requirements

app.py

@@ -0,0 +1,46 @@
+import torch
+from PIL import Image
+import gradio as gr
+from open_clip import create_model_from_pretrained, get_tokenizer
+
+MODEL_NAME = "microsoft/BiomedCLIP-PubMedBERT_256-vit_base_patch16_224"
+
+print("Loading BiomedCLIP...")
+model, preprocess = create_model_from_pretrained(f"hf-hub:{MODEL_NAME}")
+tokenizer = get_tokenizer(f"hf-hub:{MODEL_NAME}")
+model.eval()
+print("Model Loaded")
+
+labels = [
+    "eczema",
+    "psoriasis",
+    "fungal infection",
+    "acne",
+    "dermatitis",
+    "healthy skin"
+]
+
+def predict(image):
+    image = preprocess(image).unsqueeze(0)
+
+    with torch.no_grad():
+        image_features = model.encode_image(image)
+        text_tokens = tokenizer(labels)
+        text_features = model.encode_text(text_tokens)
+
+        image_features /= image_features.norm(dim=-1, keepdim=True)
+        text_features /= text_features.norm(dim=-1, keepdim=True)
+
+        similarity = (100.0 * image_features @ text_features.T).softmax(dim=-1)
+        probs = similarity.squeeze().tolist()
+
+    results = {labels[i]: probs[i] for i in range(len(labels))}
+    return results
+
+gr.Interface(
+    fn=predict,
+    inputs=gr.Image(type="pil"),
+    outputs=gr.Label(num_top_classes=3),
+    title="BiomedCLIP Dermatology Demo",
+    description="Upload a skin image. Zero-shot prediction using BiomedCLIP. Research use only — not medical advice."
+).launch()

requirements.txt

@@ -0,0 +1,6 @@
+torch
+torchvision
+torchaudio
+open_clip_torch
+Pillow
+gradio