mirror of https://github.com/THUDM/ChatGLM-6B
Merge 6ba39b421c
into 401bf3a8a7
commit
f0f9113fa5
5
api.py
5
api.py
|
@ -2,6 +2,7 @@ from fastapi import FastAPI, Request
|
||||||
from transformers import AutoTokenizer, AutoModel
|
from transformers import AutoTokenizer, AutoModel
|
||||||
import uvicorn, json, datetime
|
import uvicorn, json, datetime
|
||||||
import torch
|
import torch
|
||||||
|
import asyncio
|
||||||
|
|
||||||
DEVICE = "cuda"
|
DEVICE = "cuda"
|
||||||
DEVICE_ID = "0"
|
DEVICE_ID = "0"
|
||||||
|
@ -29,6 +30,8 @@ async def create_item(request: Request):
|
||||||
max_length = json_post_list.get('max_length')
|
max_length = json_post_list.get('max_length')
|
||||||
top_p = json_post_list.get('top_p')
|
top_p = json_post_list.get('top_p')
|
||||||
temperature = json_post_list.get('temperature')
|
temperature = json_post_list.get('temperature')
|
||||||
|
|
||||||
|
def _sync_chat(history):
|
||||||
response, history = model.chat(tokenizer,
|
response, history = model.chat(tokenizer,
|
||||||
prompt,
|
prompt,
|
||||||
history=history,
|
history=history,
|
||||||
|
@ -48,6 +51,8 @@ async def create_item(request: Request):
|
||||||
torch_gc()
|
torch_gc()
|
||||||
return answer
|
return answer
|
||||||
|
|
||||||
|
return await asyncio.to_thread(_sync_chat, history=history)
|
||||||
|
|
||||||
|
|
||||||
if __name__ == '__main__':
|
if __name__ == '__main__':
|
||||||
tokenizer = AutoTokenizer.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True)
|
tokenizer = AutoTokenizer.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True)
|
||||||
|
|
Loading…
Reference in New Issue