Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,11 +1,12 @@
|
|
| 1 |
-
from fastapi import FastAPI
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
|
| 5 |
from llama_cpp import Llama
|
| 6 |
import time
|
| 7 |
import os
|
| 8 |
|
|
|
|
|
|
|
| 9 |
llm = Llama(model_path="llama-2-7b-chat.Q3_K_S.gguf", n_ctx=2048, n_batch=512, use_mlock=True, n_threads=8)
|
| 10 |
time.sleep(8)
|
| 11 |
os.system("cls")
|
|
@@ -13,8 +14,7 @@ print("Chatbot by Aritra Roy & DVLH")
|
|
| 13 |
import warnings
|
| 14 |
warnings.filterwarnings("ignore")
|
| 15 |
|
| 16 |
-
|
| 17 |
-
async def chat_endpoint(ask: str = Form(...)):
|
| 18 |
prompt = f"Llama-2-Chat [INST] <<SYS>>You're an assistant named Tusti. You are developed by Aritra Roy. Don't share any false information.<</SYS>> {ask} [/INST]"
|
| 19 |
|
| 20 |
output_stream = llm(prompt, max_tokens=1024, echo=False, temperature=0.2, top_p=0.1, stream=True)
|
|
@@ -25,13 +25,37 @@ async def chat_endpoint(ask: str = Form(...)):
|
|
| 25 |
chunk = next(output_stream)
|
| 26 |
if chunk.get('choices') and chunk['choices'][0].get('text'):
|
| 27 |
response_text_chunk = chunk['choices'][0]['text']
|
| 28 |
-
|
| 29 |
-
|
| 30 |
except StopIteration:
|
| 31 |
break
|
|
|
|
| 32 |
except StopIteration:
|
| 33 |
pass
|
| 34 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 35 |
if __name__ == "__main__":
|
| 36 |
import uvicorn
|
| 37 |
-
|
|
|
|
|
|
| 1 |
+
from fastapi import FastAPI
|
| 2 |
+
from fastapi.responses import StreamingResponse
|
| 3 |
+
from starlette.responses import HTMLResponse
|
|
|
|
| 4 |
from llama_cpp import Llama
|
| 5 |
import time
|
| 6 |
import os
|
| 7 |
|
| 8 |
+
app = FastAPI()
|
| 9 |
+
|
| 10 |
llm = Llama(model_path="llama-2-7b-chat.Q3_K_S.gguf", n_ctx=2048, n_batch=512, use_mlock=True, n_threads=8)
|
| 11 |
time.sleep(8)
|
| 12 |
os.system("cls")
|
|
|
|
| 14 |
import warnings
|
| 15 |
warnings.filterwarnings("ignore")
|
| 16 |
|
| 17 |
+
def generate_responses(ask):
|
|
|
|
| 18 |
prompt = f"Llama-2-Chat [INST] <<SYS>>You're an assistant named Tusti. You are developed by Aritra Roy. Don't share any false information.<</SYS>> {ask} [/INST]"
|
| 19 |
|
| 20 |
output_stream = llm(prompt, max_tokens=1024, echo=False, temperature=0.2, top_p=0.1, stream=True)
|
|
|
|
| 25 |
chunk = next(output_stream)
|
| 26 |
if chunk.get('choices') and chunk['choices'][0].get('text'):
|
| 27 |
response_text_chunk = chunk['choices'][0]['text']
|
| 28 |
+
yield response_text_chunk
|
| 29 |
+
|
| 30 |
except StopIteration:
|
| 31 |
break
|
| 32 |
+
|
| 33 |
except StopIteration:
|
| 34 |
pass
|
| 35 |
|
| 36 |
+
@app.get("/", response_class=HTMLResponse)
|
| 37 |
+
def home():
|
| 38 |
+
return """
|
| 39 |
+
<html>
|
| 40 |
+
<head>
|
| 41 |
+
<title>Chatbot Streaming</title>
|
| 42 |
+
</head>
|
| 43 |
+
<body>
|
| 44 |
+
<h1>Chatbot Streaming</h1>
|
| 45 |
+
<form action="/ask/" method="post">
|
| 46 |
+
<label for="ask">You:</label>
|
| 47 |
+
<input type="text" id="ask" name="ask" required>
|
| 48 |
+
<button type="submit">Ask</button>
|
| 49 |
+
</form>
|
| 50 |
+
</body>
|
| 51 |
+
</html>
|
| 52 |
+
"""
|
| 53 |
+
|
| 54 |
+
@app.post("/ask/")
|
| 55 |
+
def ask_endpoint(ask: str):
|
| 56 |
+
return StreamingResponse(generate_responses(ask))
|
| 57 |
+
|
| 58 |
if __name__ == "__main__":
|
| 59 |
import uvicorn
|
| 60 |
+
|
| 61 |
+
uvicorn.run(app, host="0.0.0.0", port=7856)
|