api.py

from fastapi import FastAPI, Request
from transformers import AutoTokenizer, AutoModel
import uvicorn, json, datetime
import torch
from sse_starlette.sse import EventSourceResponse

DEVICE = "cuda"
DEVICE_ID = "0"
CUDA_DEVICE = f"{DEVICE}:{DEVICE_ID}" if DEVICE_ID else DEVICE


def torch_gc():
    if torch.cuda.is_available():
        with torch.cuda.device(CUDA_DEVICE):
            torch.cuda.empty_cache()
            torch.cuda.ipc_collect()


app = FastAPI()


def predict_stream(tokenizer, prompt, history, max_length, top_p, temperature):
    for response, history in model.stream_chat(tokenizer, prompt, history, max_length=max_length, top_p=top_p,
                                               temperature=temperature):
        now = datetime.datetime.now()
        time = now.strftime("%Y-%m-%d %H:%M:%S")
        yield json.dumps({
            'response': response,
            'history': history,
            'status': 200,
            'time': time
        })
    log = "[" + time + "] " + '", prompt:"' + prompt + '", response:"' + repr(response) + '"'
    print(log)
    return torch_gc()


def predict(tokenizer, prompt, history, max_length, top_p, temperature):
    response, history = model.chat(tokenizer,
                                   prompt,
                                   history=history,
                                   max_length=max_length if max_length else 2048,
                                   top_p=top_p if top_p else 0.7,
                                   temperature=temperature if temperature else 0.95)
    now = datetime.datetime.now()
    time = now.strftime("%Y-%m-%d %H:%M:%S")
    answer = {
        "response": response,
        "history": history,
        "status": 200,
        "time": time
    }
    log = "[" + time + "] " + '", prompt:"' + prompt + '", response:"' + repr(response) + '"'
    print(log)
    torch_gc()
    return answer

@app.post("/")
async def create_item(request: Request):
    global model, tokenizer
    json_post_raw = await request.json()
    json_post = json.dumps(json_post_raw)
    json_post_list = json.loads(json_post)
    prompt = json_post_list.get('prompt')
    history = json_post_list.get('history')
    max_length = json_post_list.get('max_length')
    top_p = json_post_list.get('top_p')
    temperature = json_post_list.get('temperature')
    stream = json_post_list.get('stream')
    if stream:
        res = predict_stream(tokenizer, prompt, history, max_length, top_p, temperature)
        return EventSourceResponse(res)
    else:
        answer = predict(tokenizer, prompt, history, max_length, top_p, temperature)
        return answer


if __name__ == '__main__':
    tokenizer = AutoTokenizer.from_pretrained("model", trust_remote_code=True)
    model = AutoModel.from_pretrained("model", trust_remote_code=True).half().cuda()
    model.eval()
    uvicorn.run(app, host='192.168.3.86', port=7860, workers=1)