Update app.py

app.py

@@ -22,7 +22,7 @@ MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
 HF_TOKEN = os.getenv("HF_TOKEN")
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
-model_name = "datalama/EXAONE-3.5-7.8B-Instruct-Llamafied"
+model_name = "ehristoforu/Phi4-MoE-2x14B-Instruct"
 
 model = AutoModelForCausalLM.from_pretrained(
     model_name,
@@ -34,9 +34,9 @@ tokenizer = AutoTokenizer.from_pretrained(model_name)
 #peft_model = AutoPeftModelForCausalLM.from_pretrained("ehristoforu/CoolQwen2.5-3b-it")
 #merged_model = peft_model.merge_and_unload()
 #merged_model.save_pretrained("./coolqwen")
-#model.save_pretrained("./coolqwen")
-#tokenizer.save_pretrained("./coolqwen")
-'''
+model.save_pretrained("./coolqwen")
+tokenizer.save_pretrained("./coolqwen")
+
 from huggingface_hub import HfApi
 
 api = HfApi()
@@ -45,11 +45,11 @@ api = HfApi()
 
 api.upload_folder(
     folder_path="./coolqwen",
-    repo_id="ehristoforu/coolqwen-3b-it",
+    repo_id="ehristoforu/Phi4-MoE-2x14B",
     repo_type="model",
     token=HF_TOKEN,
 )
-'''
+
 
 @spaces.GPU(duration=60)
 def generate(