forked from Tele-AI/TeleChat-52B
-
Notifications
You must be signed in to change notification settings - Fork 0
/
infer.py
27 lines (25 loc) · 1.52 KB
/
infer.py
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
import os
import torch
import pandas as pd
from transformers import AutoModelForCausalLM, AutoTokenizer
from transformers import GenerationConfig
os.environ["CUDA_VISIBLE_DEVICES"]="0,1,2,3"
PATH = "/PATH/TO/Models" ## 模型路徑
tokenizer = AutoTokenizer.from_pretrained(PATH, use_fast=False, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(PATH,
torch_dtype=torch.bfloat16,
device_map='auto',
trust_remote_code=True)
model.eval()
if __name__ == '__main__':
questions = ["世界上最潮湿的地方是哪里?",
"你作为一名气候保护协会的会员,你准备写一篇全球气候变化的新闻报告,要求体现出全球气候变化以前与现在情况的对比,字数要求1000字。"]
#### single turn example
generate_config = GenerationConfig.from_pretrained(PATH)
for question in questions:
answer = model.chat(tokenizer,question, history_input_list = [], history_output_list = [],generation_config = generate_config)
print("machine:",answer)
### multi turn example
for question in questions:
answer = model.chat(tokenizer,question, history_input_list=["愚人节是在每一年的哪一天?","世界贸易组织的缩写是?"], history_output_list=["愚人节是每年的4月1日。","世界贸易组织(WTO)的缩写是WTO。"],generation_config = generate_config)
print("machine:",answer)