Fix chat_with_model

app.py

@@ -102,6 +102,12 @@ def chat_with_model(messages):
         yield messages
         return
 
+    # 🛠 Missing variable initializations
+    max_new_tokens = 1024
+    output_text = ""
+    in_think = False
+    generated_tokens = 0
+
     pad_id = current_tokenizer.pad_token_id
     eos_id = current_tokenizer.eos_token_id
     if pad_id is None:
@@ -137,7 +143,6 @@ def chat_with_model(messages):
 
     full_messages = system_messages + filtered_messages
 
-
     prompt = format_prompt(full_messages)
 
     device = torch.device("cuda")
@@ -198,9 +203,8 @@ def chat_with_model(messages):
 
         messages[-1]["content"] = output_text
 
-        current_id = patient_id.value
-        if current_id:
-            patient_conversations[current_id] = messages
+        # Save conversation per patient
+        patient_conversations[current_id] = messages
 
         yield messages
 
@@ -213,10 +217,6 @@ def chat_with_model(messages):
     return messages
 
 
-
-
-
-
 def load_model_on_selection(model_name, progress=gr.Progress(track_tqdm=False)):
     global current_model, current_tokenizer
     token = os.getenv("HF_TOKEN")