-
因为最近在做一些阅读理解相关的工作,想试一试ernie,但我看在github上并没有在任何MRC数据集(比如DuReader)上跑出来的结果,不知道有没有人做过?
-
您好, 在这个目录下 `Dureader-Bert/Dureader/predict/predicting.py`, 有一段代码没看懂
在`find_best_answer_for_passage`函数中, `while True`跑三个循环这里是为什么这么定义的,
每次预测出的start跟end位置相反就全部置为0重来, 似乎讲不通..
感觉跟其他模型中用动态规划的思路求最大score的…
-
- 版本、环境信息:
1)GPU:CUDA版本=9,CUDNN版本=7。GPU设备为2080Ti,在TensorFlow-gpu和PyTorch-gpu上可以正常安装和使用GPU。
2)系统环境:系统=16.04.1-Ubuntu、Python版本=Python 2.7.12
安装方式信息:
1)pip安装:pip install paddlepaddle-gpu
问题描述…
-
顺便吐槽下 ai.baidu.com , 跳转不带tag,从DuReader里点击download,弹出来的页面竟然是download页面里的DuConv
不知道找谁反馈
-
我用data/download.sh下载了dureader_preprocessed.zip,解压得到testset、trainset、devset的各个json,我把每个json都head -n 3000截取头3000行。前面的步骤没有错误,到了train这步出现一堆SKIP:
![paddle-train](https://user-images.githubusercontent.com/…
-
这个tensorflow使用了CNN对character进行embedding吗?
-
When i ran the BiDAF model (origin code without any modification) in the tensorflow version, the training loss decreases to about 2. However, if we ran a dev mode after each epoch, the dev loss does …
-
-
感谢作者分享!
有个小问题请教,
MRC2018-master/BiDAF_Origin/dureader/utils/baseline_eval.py中,引入两个包
from brc_eval import compute_metrics_from_list
from find_answer import find_best_question_match
但我在代码中没发现这两个包,作…
-
mldl@mldlUB1604:~/ub16_prj/DuReader$ cat data/raw/trainset/search.train.json | python3 utils/preprocess.py > data/preprocessed/trainset/search.train.json
Traceback (most recent call last):
File "u…