max concurrenct 1 fix

agents.py

@@ -108,7 +108,7 @@ def describe_with_voice(state: State) -> State:
     input_len = inputs["input_ids"].shape[-1]
 
     with torch.inference_mode():
-        generation = model.generate(**inputs, max_new_tokens=1000, do_sample=False)
+        generation = model.generate(**inputs, max_new_tokens=1000, do_sample=True, temperature=0.7)
         generation = generation[0][input_len:]
 
     description = processor.decode(generation, skip_special_tokens=True)

app.py

@@ -13,7 +13,7 @@ def process_and_display(image, voice):
     state = {"image": image, "voice": voice, "caption": "", "description": ""}
 
     # Run the graph
-    result = graph.invoke(state, max_concurrency=1)
+    result = graph.invoke(state, {"max_concurrency" : 1})
 
     descriptions:list[str] = result["descriptions"]
     description = "\n---\n".join(descriptions)