Update app.py

app.py

@@ -21,8 +21,7 @@ from langchain_community.llms import HuggingFaceHub
 import pandas as pd
 
 # Configuración del modelo
-MODEL_NAME = "mistralai/Mixtral-8x7B-Instruct-v0.1"
-model_name = "google/gemma-2-2b"
+
 TOKEN=os.getenv('HF_TOKEN')
 subprocess.run(["huggingface-cli", "login", "--token", TOKEN, "--add-to-git-credential"])
 ######
@@ -33,15 +32,17 @@ os.environ["HF_TOKEN"] = st.secrets["HF_TOKEN"]
 # Initialize tokenizer
 @st.cache_resource
 def load_model():
-  # MODEL_NAME= "lmsys/vicuna-7b-v1.5"
+    TOKEN=os.getenv('HF_TOKEN')
+    subprocess.run(["huggingface-cli", "login", "--token", TOKEN, "--add-to-git-credential"])
+    os.environ["HF_TOKEN"] = st.secrets["HF_TOKEN"]
     MODEL_NAME = "google/gemma-2b-it"
 
     model = AutoModelForCausalLM.from_pretrained(
-        MODEL_NAME,\
+        MODEL_NAME
         # quantization_config=nf4_config, # add config
         # torch_dtype=torch.bfloat16, # save memory using float16
         # low_cpu_mem_usage=True,
-        token= TOKEN
+        # token= TOKEN
     ).to("cuda")
 
     tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)