Update app.py

app.py

@@ -4,7 +4,7 @@ os.system('CMAKE_ARGS="-DLLAMA_OPENBLAS=on" FORCE_CMAKE=1 pip install llama-cpp-
 import wget
 from llama_cpp import Llama
 import random
-url = 'https://huggingface.co/TheBloke/orca_mini_3B-GGML/resolve/main/orca-mini-3b.ggmlv3.q4_0.bin'
+url = 'https://huggingface.co/TheBloke/WizardLM-7B-uncensored-GGML/resolve/main/WizardLM-7B-uncensored.ggmlv3.q2_K.bin'
 filename = wget.download(url)
 llm2 = Llama(model_path=filename, seed=random.randint(1, 2**31))
 theme = gr.themes.Soft(
@@ -29,21 +29,9 @@ with gr.Blocks(theme=theme) as demo:
         #token1 = llm.tokenize(b"### Instruction: ")
         #token2 = llm.tokenize(instruction.encode())
         #token3 = llm2.tokenize(b"USER: ")
-        token1 = llm2.tokenize(b'''
-        ### System:
-        You are an human that follows instruction extremely well. Help as much as you can.
-        ''')
-        token2 = llm2.tokenize(b'''
-        ### User:
-        ''')
-        token3 = llm2.tokenize(user_message.encode())
-        token4 = llm2.tokenize(b'''
-        ### Response:
-        ''')
-        tokens = token1 + token2 + token3 + token4
-        #tokens3 = llm2.tokenize(user_message.encode())
-        #token4 = llm2.tokenize(b"\n\n### Response:")
-        #tokens = tokens3 + token4
+        tokens5 = llm2.tokenize(user_message.encode())
+        token4 = llm2.tokenize(b"\n\n### Response:")
+        #tokens = tokens5 + token4
         history[-1][1] = ""
         count = 0
         output = ""