PaddleNLP issues

UIE抽取耗时20ms，是否可继续缩短？

11

模型训练完毕后，UIE抽取耗时目前稳定在20ms左右。显卡为16GB，已设置fp16。是否还有其他方法可使耗时能继续缩短？

- 版本、环境信息 1）PaddleNLP 2.3，PaddlePaddle2.3 2）系统环境：Linux，python3.7 3）batch_size=1,max_seq_lenth=512,train600条，test200条，dev200条 `# 模型训练： import paddle.nn.functional as F import time save_dir = "checkpoint/bert-wwm" if not os.path.exists(save_dir): os.makedirs(save_dir) save_train_result = "./results/bert-wwm.tsv" train_r_df = pd.DataFrame(data=None, columns=["global_step","epoch","step","loss","acc","time"]) pre_accu=0 accu=0...

zoeChen119

Add text summary application

### PR types New features ### PR changes Application ### Description

gongel

Use fastdeploy instead of inference backend

### PR types New features ### PR changes Others ### Description Use fastdeploy instead of inference backend

joey12300

[NEW MODEL] Add CANINE model

6

### PR types New features ### PR changes Models ### Description 【飞桨论文复现挑战赛（第六期）】111号论文 [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://paperswithcode.com/paper/canine-pre-training-an-efficient-tokenization) 。添加 Canine 模型，复现仓库地址 [github](https://github.com/kevinng77/canine_paddle)，[Aistudio](https://aistudio.baidu.com/aistudio/projectdetail/4063353?channelType=0&channel=0)。

kevinng77

contributor

在aistudio上运行样例的时候，提示information_extraction不在Taskflow list中

2

`--------------------------------------------------------------------------- AssertionError Traceback (most recent call last) /tmp/ipykernel_182/1763430471.py in 3 4 schema = ['时间', '选手', '赛事名称'] ----> 5 ie = Taskflow('information_extraction', schema=schema) 6 pprint(ie("2月8日上午北京冬奥会自由式滑雪女子大跳台决赛中中国选手谷爱凌以188.25分获得金牌！")) /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddlenlp/taskflow/taskflow.py in __init__(self, task, model, device_id,...

estuday