Spaces:
Sleeping
Sleeping
File size: 1,439 Bytes
d809ddf 48d8d65 d809ddf ef8deae 48d8d65 |
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 |
from fastapi import FastAPI, Request
from fastapi.responses import StreamingResponse
from pydantic import BaseModel
from vllm import AsyncLLMEngine, SamplingParams
import asyncio
import json
app = FastAPI()
# Initialize the AsyncLLMEngine
# Replace 'your-model-path' with the actual path or name of your model
engine = AsyncLLMEngine.from_pretrained('microsoft/Phi-3-mini-4k-instruct')
class GenerationRequest(BaseModel):
prompt: str
max_tokens: int = 100
temperature: float = 0.7
async def generate_stream(prompt: str, max_tokens: int, temperature: float):
sampling_params = SamplingParams(
temperature=temperature,
max_tokens=max_tokens
)
async for output in engine.generate(prompt, sampling_params, True): # True enables streaming
yield f"data: {json.dumps({'text': output.outputs[0].text})}\n\n"
yield "data: [DONE]\n\n"
@app.post("/generate-stream")
async def generate_text(request: Request):
try:
data = await request.json()
gen_request = GenerationRequest(**data)
return StreamingResponse(
generate_stream(gen_request.prompt, gen_request.max_tokens, gen_request.temperature),
media_type="text/event-stream"
)
except Exception as e:
return StreamingResponse(
iter([f"data: {json.dumps({'error': str(e)})}\n\n"]),
media_type="text/event-stream"
)
|