Spaces:
Runtime error
Runtime error
File size: 1,282 Bytes
07be806 402607b 2b38ca8 402607b 07be806 402607b 07be806 3e9f81b f0b6f7d 07be806 f0b6f7d 07be806 f0b6f7d 07be806 402607b 07be806 f0b6f7d 402607b f0b6f7d 672d7aa | 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 | from fastapi import FastAPI, Request
from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
import torch
import uvicorn
app = FastAPI(title="Phi-2 Free API")
# ✅ Non-gated model (public, CPU-friendly)
model_name = "microsoft/phi-2"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.float32)
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=200,
temperature=0.7,
top_p=0.9,
repetition_penalty=1.1,
)
@app.get("/")
async def home():
return {"message": "✅ Phi-2 API is running!"}
@app.post("/chat")
async def chat(request: Request):
data = await request.json()
prompt = data.get("prompt", "").strip()
if not prompt:
return {"error": "Please provide a 'prompt' in JSON."}
# Generate text
result = pipe(prompt, max_new_tokens=150, do_sample=True)[0]["generated_text"]
# ✂️ Clean response: remove prompt echo
cleaned = result.replace(prompt, "").strip()
if not cleaned:
cleaned = "(No response generated. Try a longer prompt or lower temperature.)"
return {"response": cleaned}
if __name__ == "__main__":
uvicorn.run(app, host="0.0.0.0", port=7860)
|