File size: 3,412 Bytes
20a070f 6eb0c45 20a070f bed5475 20a070f bed5475 20a070f |
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 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 |
import argparse
import asyncio
import functools
import json
import os
from io import BytesIO
import uvicorn
from fastapi import FastAPI, BackgroundTasks, File, Body, UploadFile, Request
from fastapi.responses import StreamingResponse
from starlette.staticfiles import StaticFiles
from starlette.templating import Jinja2Templates
from sentence_transformers import SentenceTransformer
# from utils.data_utils import remove_punctuation
# from utils.utils import add_arguments, print_arguments
def print_arguments(args):
print("----------- Configuration Arguments -----------")
for arg, value in vars(args).items():
print("%s: %s" % (arg, value))
print("------------------------------------------------")
def strtobool(val):
val = val.lower()
if val in ('y', 'yes', 't', 'true', 'on', '1'):
return True
elif val in ('n', 'no', 'f', 'false', 'off', '0'):
return False
else:
raise ValueError("invalid truth value %r" % (val,))
def str_none(val):
if val == 'None':
return None
else:
return val
def add_arguments(argname, type, default, help, argparser, **kwargs):
type = strtobool if type == bool else type
type = str_none if type == str else type
argparser.add_argument("--" + argname,
default=default,
type=type,
help=help + ' Default: %(default)s.',
**kwargs)
os.environ['KMP_DUPLICATE_LIB_OK'] = 'True'
parser = argparse.ArgumentParser(description=__doc__)
add_arg = functools.partial(add_arguments, argparser=parser)
add_arg("host", type=str, default="0.0.0.0", help="")
add_arg("port", type=int, default=5000, help="")
add_arg("model_path", type=str, default="BAAI/bge-small-en-v1.5", help="")
add_arg("use_gpu", type=bool, default=False, help="")
# add_arg("use_int8", type=bool, default=True, help="")
add_arg("beam_size", type=int, default=10, help="")
add_arg("num_workers", type=int, default=2, help="")
add_arg("vad_filter", type=bool, default=True, help="")
add_arg("local_files_only", type=bool, default=True, help="")
args = parser.parse_args()
print_arguments(args)
#
assert os.path.exists(args.model_path), f"{args.model_path}"
#
if args.use_gpu:
model = SentenceTransformer(args.model_path, device="cuda", compute_type="float16")
else:
model = SentenceTransformer(args.model_path, device='cpu')
app = FastAPI(title="embedding Inference")
# app.mount('/static', StaticFiles(directory='static'), name='static')
# templates = Jinja2Templates(directory="templates")
# model_semaphore = None
@app.post("/embed")
async def api_embed(
textA: str = Body("text1", description="", embed=True),
textB: str = Body("text2", description="", embed=True),
):
q_embeddings = model.encode(textA, normalize_embeddings=True)
p_embeddings = model.encode(textB, normalize_embeddings=True)
scores = q_embeddings @ p_embeddings.T
print(scores)
scores = scores.tolist()
ret = {"similarity score": scores, "status_code": 200}
return ret
# @app.get("/")
# async def index(request: Request):
# return templates.TemplateResponse(
# "index.html", {"request": request, "id": id}
# )
if __name__ == '__main__':
uvicorn.run(app, host=args.host, port=args.port) |