Update app.py

app.py

@@ -1,25 +1,22 @@
-import subprocess
 from threading import Thread
 from llama_cpp import Llama
 import torch
 import gradio as gr
 import re
 
-subprocess.call(f"""CMAKE_ARGS="-DLLAMA_CUBLAS=on" FORCE_CMAKE=1 pip install --upgrade --force-reinstall llama-cpp-python --no-cache-dir
-""", shell=True)
 
 torch_device = "cuda" if torch.cuda.is_available() else "cpu"
 print("Running on device:", torch_device)
 print("CPU threads:", torch.get_num_threads())
 
 llm = Llama(model_path = 'Llama-2-ko-7B-chat-gguf-q4_0.bin',
-            n_ctx=50,
+            n_ctx=40,
             n_threads = 8,
             n_batch = 5            
       )
 
 def gen(x, max_new_tokens):
-    output = llm(f"### 명령어:\n{x}\n\n### 응답:", max_tokens=max_new_tokens, stop=["###", "\n", ":"], echo=True)
+    output = llm(f"### 명령어:\n{x}\n\n### 응답:", max_tokens=max_new_tokens, stop=["###"], echo=True)
     
     return output['choices'][0]['text'].replace('▁',' ')