Update app.py

app.py

@@ -13,11 +13,11 @@ DEFAULT_MAX_NEW_TOKENS = 512
 MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
 
 #Inference API Code
-client = InferenceClient("BenBranyon/zephyr-sumbot-all-songs-large")
+#client = InferenceClient("BenBranyon/zephyr-sumbot-all-songs-large")
 
 #Transformers Code
 if torch.cuda.is_available():
-    model_id = "BenBranyon/zephyr-sumbot-all-songs"
+    model_id = "BenBranyon/zephyr-sumbot-all-songs-large"
     model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")
     tokenizer = AutoTokenizer.from_pretrained(model_id)
     tokenizer.use_default_system_prompt = False