Change of settings again for TPU with 6bit Model

app.py

@@ -47,7 +47,7 @@ REPO_ID = "sugiv/leetmonkey-peft-gguf"
 model_path = hf_hub_download(repo_id=REPO_ID, filename=MODEL_NAME, cache_dir="./models")
 #llm = Llama(model_path=model_path, n_ctx=2048, n_threads=16, n_gpu_layers=-1, verbose=False, mlock=True) ## TPU
 #llm = Llama(model_path=model_path, n_ctx=1024, n_threads=2, n_gpu_layers=0, verbose=False, mlock=False) ## CPU only
-llm = Llama(model_path=model_path, n_ctx=1024, n_threads=4, n_gpu_layers=10, verbose=False, mlock=False) ## Nvidia
+llm = Llama(model_path=model_path, n_ctx=1024, n_threads=4, n_gpu_layers=-1, verbose=False, mlock=True) ## Nvidia
 logger.info("8-bit model loaded successfully")
 
 # User data storage