add qwen inference

backend/src/agents/expert.py

@@ -52,9 +52,9 @@ class ExpertAgent(Agent):
         model = self.config['model']
 
         if model == "qwen-big":
-            response = await inference.replicate_qwen_completion(self.app, messages, model=inference.Models.quen3_235b)
+            response = await inference.replicate_qwen_completion(self.app, messages, model=inference.Models.qwen3_235b)
         elif model == "qwen-small":
-            response = await inference.replicate_qwen_completion(self.app, messages, model=inference.Models.quen3_8b)
+            response = await inference.replicate_qwen_completion(self.app, messages, model=inference.Models.qwen3_8b)
         elif model == "llama-big":
             response = await inference.replicate_llama_completion(self.app, messages, model=inference.Models.llama_70b)
         elif model == "llama-small":

backend/src/inference.py

@@ -15,7 +15,7 @@ class Models:
     llama_8b = "meta/meta-llama-3-8b-instruct"
     llama_70b = "meta/meta-llama-3-70b-instruct"
     qwen3_8b = "lucataco/qwen3-vl-8b-instruct:39e893666996acf464cff75688ad49ac95ef54e9f1c688fbc677330acc478e11"
-    quen3_235b = "qwen/qwen3-235b-a22b-instruct-2507"
+    qwen3_235b = "qwen/qwen3-235b-a22b-instruct-2507"
 
 async def replicate_llama_completion(app, messages, model=None, opts={}):
     if model is None:
@@ -39,7 +39,7 @@ async def replicate_qwen_completion(app, messages, model=None, opts={}):
             "temperature": opts.get('temperature', 0.75),
             "max_new_tokens": opts.get('max_tokens', 2048),
         }
-    elif model == Models.quen3_235b:
+    elif model == Models.qwen3_235b:
         input = {
             "prompt": format_instruct(messages),
             "temperature": opts.get('temperature', 0.75),