Add HF_TOKEN auth + fallback model for robustness

app.py

@@ -5,9 +5,25 @@ import os
 import gradio as gr
 from huggingface_hub import InferenceClient
 
-# Models available on free serverless inference (no token needed for these)
-# Qwen2.5-72B-Instruct is available on serverless - use it for best quality
-client = InferenceClient("Qwen/Qwen2.5-72B-Instruct")
+HF_TOKEN = os.environ.get("HF_TOKEN")
+
+# Try multiple providers in order of preference
+def get_client():
+    models = [
+        "Qwen/Qwen2.5-72B-Instruct",
+        "Qwen/Qwen2.5-7B-Instruct",
+        "meta-llama/Llama-3.1-8B-Instruct",
+    ]
+    for model in models:
+        try:
+            c = InferenceClient(model, token=HF_TOKEN)
+            # Quick test
+            return c, model
+        except Exception:
+            continue
+    return InferenceClient(models[0], token=HF_TOKEN), models[0]
+
+client, MODEL_USED = get_client()
 
 SYSTEM_PROMPT = """You are DeepMed-R1, a medical reasoning AI trained with GRPO and multi-objective clinical rewards on AMD MI300X.
 
@@ -21,13 +37,13 @@ For every medical question, demonstrate systematic clinical reasoning:
 
 Present reasoning inside <think></think> tags, then provide your final answer.
 For MCQ, end with \\boxed{X} where X is the correct letter.
-Be thorough but concise. Always ground reasoning in pathophysiology and evidence."""
+Be thorough but concise. Ground reasoning in pathophysiology and evidence."""
 
 EXAMPLES = [
     ["A 65-year-old male with hypertension presents with sudden 'worst headache of my life,' neck stiffness, photophobia. BP 180/100. Most likely diagnosis?\nA. Migraine\nB. Subarachnoid hemorrhage\nC. Meningitis\nD. Tension headache"],
     ["A 28-year-old woman: fatigue, weight gain, cold intolerance. TSH 12 mIU/L, Free T4 0.5 ng/dL. Initial treatment?\nA. Levothyroxine\nB. Liothyronine\nC. Methimazole\nD. Radioactive iodine"],
     ["3-month-old with projectile non-bilious vomiting, olive-shaped RUQ mass, metabolic alkalosis. Diagnosis?\nA. Pyloric stenosis\nB. Intussusception\nC. Malrotation\nD. Hirschsprung disease"],
-    ["55-year-old diabetic: RUQ pain, fever 39.2C, jaundice (Charcot's triad). WBC 18K, bilirubin 5.2, CBD stone on US. Next step?\nA. Cholecystectomy\nB. ERCP with sphincterotomy\nC. MRCP\nD. PTC"],
+    ["55-year-old diabetic: RUQ pain, fever 39.2C, jaundice (Charcot triad). WBC 18K, bilirubin 5.2, CBD stone. Next step?\nA. Cholecystectomy\nB. ERCP with sphincterotomy\nC. MRCP\nD. PTC"],
     ["22-year-old post-MVC: left chest pain, absent breath sounds left, trachea deviated right, JVD, BP 80/50. Immediate management?\nA. Chest X-ray\nB. CT chest\nC. Needle decompression left chest\nD. Intubation"],
 ]
 
@@ -46,11 +62,19 @@ def respond(message, history):
             response += delta
             yield response
     except Exception as e:
-        yield f"Error: {str(e)}\n\nPlease try again or check if the model is available."
+        error_msg = str(e)
+        if "api_key" in error_msg.lower() or "token" in error_msg.lower() or "401" in error_msg:
+            yield ("⚠️ **Authentication Required**\n\n"
+                   "Please add your HF_TOKEN as a Space secret:\n"
+                   "1. Go to Space Settings → Repository secrets\n"
+                   "2. Add secret: Name=`HF_TOKEN`, Value=your token\n\n"
+                   f"Error: {error_msg}")
+        else:
+            yield f"Error: {error_msg}\n\nPlease try again."
 
 
 with gr.Blocks(title="DeepMed-R1", theme=gr.themes.Soft(primary_hue="blue")) as demo:
-    gr.Markdown("""
+    gr.Markdown(f"""
 # 🏥 DeepMed-R1: Medical Reasoning AI
 
 **Systematic clinical reasoning powered by GRPO + Multi-Objective Clinical Rewards**
@@ -77,9 +101,9 @@ Built for **AMD Developer Hackathon 2026** — Track 2: Fine-Tuning on AMD GPUs
 | Training | GRPO + DAPO loss + CRPO rewards |
 | Innovations | iGRPO (Feb 2026) + AERO (Feb 2026) + Curriculum (Mar 2026) |
 | Hardware | AMD MI300X (192GB HBM3) |
-| Reward System | Accuracy (w=2.0) + Reasoning (w=1.0) + Consistency (w=0.5) + Length (w=0.3) |
+| Reward | Accuracy (2.0) + Reasoning (1.0) + Consistency (0.5) + Length (0.3) |
 
-📄 [Code & Training Pipeline](https://huggingface.co/asusf15/DeepMed-R1) |
+📄 [Code & Training](https://huggingface.co/asusf15/DeepMed-R1) |
 📚 [Gazal-R1](https://arxiv.org/abs/2506.21594) |
 📚 [Clinical-R1](https://arxiv.org/abs/2512.00601) |
 📚 [iGRPO](https://arxiv.org/abs/2602.09000)