Update app.py

app.py

@@ -16,11 +16,11 @@ subprocess.run(
 
 DESCRIPTION = '''
 <div>
-<h1 style="text-align: center;">Lexora-Medium-7B</h1>
-<p>This Space demonstrates the instruction-tuned model <a href="https://huggingface.co/DeepMount00/Lexora-Medium-7B"><b>Lexora-Medium-7B Chat ITA</b></a>.</p>
+<h1 style="text-align: center;">Lexora-Medium-3B</h1>
+<p>This Space demonstrates the instruction-tuned model <a href="https://huggingface.co/DeepMount00/Lexora-Lite-3B"><b>Lexora-Lite-3B Chat ITA</b></a>.</p>
 </div>
 <div>
-  <p>This model, <strong>DeepMount00/Lexora-Medium-7B</strong>, is currently the best open-source large language model for the Italian language. You can view its ranking and compare it with other models on the leaderboard at <a href="https://huggingface.co/spaces/FinancialSupport/open_ita_llm_leaderboard"><b>this site</b></a>.</p>
+  <p>This model, <strong>DeepMount00/Lexora-Lite-3B</strong>, is currently the best open-source large language model for the Italian language. You can view its ranking and compare it with other models on the leaderboard at <a href="https://huggingface.co/spaces/FinancialSupport/open_ita_llm_leaderboard"><b>this site</b></a>.</p>
 </div>
 '''
 MAX_MAX_NEW_TOKENS = 2048
@@ -29,7 +29,7 @@ MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
 
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
-model_id = "DeepMount00/Lexora-Medium-7B"
+model_id = "DeepMount00/Lexora-Lite-3B"
 tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True,)
 model = AutoModelForCausalLM.from_pretrained(
     model_id,