removes truncation in tokenizer

app.py

@@ -67,7 +67,7 @@ def run_llm(prompt: str, max_new_tokens: int = 64, temperature: float = 0.0, mod
         tok, mdl = load_llm(model_id)
         
         # Tokenize input
-        inputs = tok(prompt, return_tensors="pt", truncation=True, max_length=2048)
+        inputs = tok(prompt, return_tensors="pt")
         inputs = {k: v.to(next(mdl.parameters()).device) for k, v in inputs.items()}
         
         # Generate