Update app.py

app.py

@@ -10,17 +10,20 @@ prompts = PromptLoader.load_prompts("prompts.yaml")
 
 # Available models and their configurations
 MODELS = {
-    "Zephyr 7B Beta": {
-        "name": "HuggingFaceH4/zephyr-7b-beta",
-        "provider": "huggingface"
+    "Llama 3.3 70B Instruct": {
+        "name": "meta/llama-3.3-70b-instruct",
     },
-    "Mistral 7B": {
-        "name": "mistralai/Mistral-7B-v0.1",
-        "provider": "huggingface"
+    "Llama 3.1 405B Instruct": {
+        "name": "meta/llama-3.1-405b-instruct",
     },
-    "GPT-3.5 Turbo": {
-        "name": "gpt-3.5-turbo",
-        "provider": "openai"
+    "Llama 3.2 3B Instruct": {
+        "name": "meta/llama-3.2-3b-instruct",
+    },
+    "Falcon 3 7B Instruct": {
+        "name": "tiiuae/falcon3-7b-instruct",
+    },
+    "Granite 3.0 8B Instruct": {
+        "name": "ibm/granite-3.0-8b-instruct",
     }
 }
 
@@ -34,11 +37,10 @@ PROMPT_STRATEGIES = {
 }
 
 def create_assistant(model_name):
-    model_info = MODELS[model_name]
-    if model_info["provider"] == "huggingface":
-        client = InferenceClient(model_info["name"])
-    else:  # OpenAI
-        client = OpenAI()
+
+    client = OpenAI(
+        base_url = "https://integrate.api.nvidia.com/v1",
+    )
     
     return AIAssistant(
         client=client,