Spaces:
Running
Running
File size: 1,622 Bytes
b879e43 f0f547d 3320b3e 967e0b2 48fde5c 3320b3e 7d837d2 2142f68 f0f547d f043989 2142f68 f0f547d 48fde5c 2142f68 3320b3e 2142f68 ad138c6 2257bba 2142f68 3320b3e 2142f68 00a3720 3320b3e | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 | from fastapi import FastAPI, Request, HTTPException, Depends
from fastapi.middleware.cors import CORSMiddleware
from llama_cpp import Llama
import uvicorn
app = FastAPI()
# --- CORS SETTINGS (Crucial for GitHub Pages) ---
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# --- CONFIGURATION ---
MY_API_KEY = "my-secret-key-456"
# Optimized for HF Free Tier CPU
llm = Llama(model_path="./model.gguf", n_ctx=2048, n_threads=4, n_gpu_layers=0)
def verify_key(request: Request):
auth = request.headers.get("Authorization")
if auth != f"Bearer {MY_API_KEY}":
raise HTTPException(status_code=403, detail="Unauthorized")
@app.get("/")
def home():
return {"status": "Online", "branding": "ChatMPT Team"}
@app.post("/v1/chat")
async def chat(request: Request, _ = Depends(verify_key)):
body = await request.json()
user_input = body.get("prompt", "")
# Instruction-based prompt to fix the identity
prompt = (
"Assistant is a polite AI named ChatMPT, created by the ChatMPT Team.\n"
f"User: {user_input}\n"
"Assistant:"
)
response = llm(
prompt,
max_tokens=500,
stop=["User:", "\n", "Assistant:"],
temperature=0.7
)
raw_reply = response["choices"][0]["text"].strip()
# Anti-Typo Safety Filter
final_reply = raw_reply.replace("ChatPBT", "ChatMPT").replace("ChatPP", "ChatMPT")
return {"reply": final_reply}
if __name__ == "__main__":
uvicorn.run(app, host="0.0.0.0", port=7860) |