Create model.py

model.py

@@ -0,0 +1,23 @@
+import os
+import torch
+from huggingface_hub import login
+from transformers import AutoProcessor, AutoModelForImageTextToText
+
+# Login with your secret token
+login(token=os.environ["HF_TOKEN"])
+
+MODEL_ID = "google/medgemma-1.5-4b-it"
+
+print("Loading MedGemma... this may take a few minutes")
+
+processor = AutoProcessor.from_pretrained(MODEL_ID)
+
+model = AutoModelForImageTextToText.from_pretrained(
+    MODEL_ID,
+    torch_dtype=torch.bfloat16,
+    device_map="auto"
+)
+
+model.eval()
+
+print("MedGemma loaded successfully!")