Update app.py

app.py

@@ -166,10 +166,10 @@ print("Running LLM Mistral")
 llm_mistral = Llama(model_path=mistral_model_path,n_gpu_layers=GPU_LAYERS,max_new_tokens=256, context_window=4096, n_ctx=4096,n_batch=128,verbose=LLAMA_VERBOSE)
 
 print("Running LLM Zephyr")
-llm_zephyr = Llama(model_path=zephyr_model_path,n_gpu_layers=GPU_LAYERS-5,max_new_tokens=256, context_window=4096, n_ctx=4096,n_batch=128,verbose=LLAMA_VERBOSE)
+llm_zephyr = Llama(model_path=zephyr_model_path,n_gpu_layers=GPU_LAYERS-15,max_new_tokens=256, context_window=4096, n_ctx=4096,n_batch=128,verbose=LLAMA_VERBOSE)
 
 print("Running Yi LLM")
-llm_zephyr = Llama(model_path=yi_model_path,n_gpu_layers=GPU_LAYERS-5,max_new_tokens=256, context_window=4096, n_ctx=4096,n_batch=128,verbose=LLAMA_VERBOSE)
+llm_zephyr = Llama(model_path=yi_model_path,n_gpu_layers=GPU_LAYERS-15,max_new_tokens=256, context_window=4096, n_ctx=4096,n_batch=128,verbose=LLAMA_VERBOSE)
 
 
 # Mistral formatter