test/fastNLP
1
0
Fork 0
mirror of https://gitee.com/fastnlp/fastNLP.git synced 2024-12-04 13:17:51 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity
1,767 Commits 3 Branches 13 Tags 25 MiB
5ed688b3bc
Commit Graph

1767 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
ChenXin
2d1500651f 三个纬度 check: DataSet Tutorial 2020-02-27 21:21:38 +08:00
ChenXin
72eb6b95c2 消失的代码块修复【 code-block:: 和 python 之间需要一个空格 】 2020-02-27 21:21:32 +08:00
ChenXin
cee6b7fb45 将序列标注问题移到快速入门 2020-02-27 21:20:57 +08:00
yh_cc
9bed203a35 Merge branch 'master' of github.com:fastnlp/fastNLP 2020-02-20 17:53:15 +08:00
yh_cc
51e8c3c3b4 update static_embedding to make its property words_to_words is sequential 2020-02-20 17:53:09 +08:00
ROGERDJQ
f3ee16a5f6
[new] add ConfusionMatrix, ConfusionMatrixMetric (#272) 
* add ConfusionMatrix, ConfusionMatrixMetric

* add confusionmatrix to utils

* add ConfusionMatrixmetric

* add ConfusionMatrixMetric

* init for test

* begin test

* test finish

* doc finish
 2020-02-13 20:38:10 +08:00
Yige Xu
58762810ec [update] Re-organize classification loader and pipe 2020-02-08 17:16:24 +08:00
Yige Xu
0aab5d5277 [bugfix] fix an warning error in MNLI loader 2020-01-26 23:18:03 +08:00
Yige Xu
380e88273d [bugfix] delete a redundant import statement in io/pipe/classification.py 2020-01-12 10:45:01 +08:00
Yige Xu
46ea42498d [update] add Loader and Pipe for AG's News dataset 2020-01-12 10:40:33 +08:00
ChenXin
5f1d0cc4ee
Update __init__.py 
Fix the wrong  "__version__"  value.
 2020-01-09 20:34:26 +08:00
yh_cc
ce6b72232a update FitlogCallback使它在使用update_every时也能正确记录loss 2019-12-28 22:59:11 +08:00
yh_cc
4518a8fccb [bug fix] test bug fix 2019-12-27 23:12:29 +08:00
yh_cc
7024df46af update SaveModelCallback和CheckPointCallback, 使他们可以在DataParallel的情况下正常工作 2019-12-27 21:32:02 +08:00
yh_cc
55aab72eb2 Merge branch 'master' of github.com:fastnlp/fastNLP 2019-12-27 21:08:28 +08:00
yunfan
1139986b36 [bugfix] fix logger's logging level 2019-12-27 19:23:44 +08:00
yh_cc
25948294c7 CheckPointCallback中recover_fitlog自动识别 2019-12-27 18:44:21 +08:00
yh_cc
9293a6c1ab 修复一个小bug 
新增CheckPointCallback用于恢复Trainer的训练; CheckPointCallback会将optimizer，model以及Trainer的状态恢复到保存的epoch
 2019-12-27 18:42:35 +08:00
yh_cc
699a0ef74d 1.增强viterbi解码使得可以加入start, end; 2.修改BertEnocder中注释的typo 2019-12-26 13:05:13 +08:00
yh_cc
218f3aaded 1.修复BertEncoder中的flip时报错的bug; 2.删除Metric中fast_param_map防止误触发导致不好debug的问题 2019-12-24 00:30:38 +08:00
yh_cc
7db26a425d Merge branch 'master' of github.com:fastnlp/fastNLP 2019-12-23 10:38:24 +08:00
yh_cc
1d3b80b7f9 修复_move_model_to_device中当device=torch.device('cpu')会触发的bug 2019-12-23 10:37:41 +08:00
Xipeng Qiu
1cf0a76a1d Update README.md 2019-12-20 11:12:59 +08:00
Xipeng Qiu
4a8d40340b update readme 2019-12-20 11:09:11 +08:00
yh_cc
09078d0250 使用Trainer中data_iterator属性参与循环使得可以通过callback修改数据 2019-12-19 23:15:14 +08:00
Yunfan Shao
1b377ef851
Merge pull request #264 from Gosicfly/master 
出现非法格式catch住的时候，应该将sample清空
 2019-12-18 13:06:19 +08:00
Gosicfly
baf009fe5c
出现非法格式catch住的时候，应该将sample清空 
出现非法格式catch住的时候，应该将sample清空，不然会一直报warning，sample也会一直累积新的内容得不到正确的更新。
 2019-12-17 17:29:22 +08:00
yh_cc
03e9ca2147 修复CWSPipe中无法读取包含<, >的数据的问题https://github.com/fastnlp/fastNLP/issues/258#issue-533768454 2019-12-13 22:19:38 +08:00
Yige Xu
28ddc0c17e [update]对最近几个issue的更新 2019-12-12 17:00:48 +08:00
Yige Xu
980aba9898 [update] update index documents 2019-11-30 05:31:06 +08:00
Yige Xu
2033e1d2d1 [uodate]更新metric的文档显示 2019-11-30 04:48:36 +08:00
yh_cc
ae8079cd56 Merge branch 'master' of github.com:fastnlp/fastNLP 2019-11-29 11:47:04 +08:00
yh_cc
ac0454c89b 更新部分内部引用 2019-11-29 11:46:57 +08:00
yh
5cbe1a61d6 Merge branch 'master' of https://github.com/fastnlp/fastNLP 2019-11-28 17:37:28 +08:00
yh
39f1185e1d 修复BiLSTM中ht, ct状态在有seq_len时顺序错乱的bug 2019-11-28 17:37:18 +08:00
Yige Xu
dd314a3d4f [docfix] 修复一个BertEmbedding时warning信息的拼写错误 2019-11-25 15:18:43 +08:00
yh
08c4bc86c7 1.修复char_embedding中char encoder在获取char embedding dimension时可能发生的bug; 2.将staticembedding中only_train_min_freq默认为False; 3. check_loader_paths()不会对重复的文件名报错，修改了该bug 2019-11-24 23:02:52 +08:00
yh_cc
fe1b47b15a Merge branch 'master' of github.com:fastnlp/fastNLP 2019-11-22 18:07:54 +08:00
yh_cc
faec3245d4 update CMRC2018Metric测试 2019-11-22 18:07:46 +08:00
yh_cc
be0b588616 update 文档中的typo 2019-11-21 17:19:08 +08:00
yunfan
03bee90521 [update] docs for logger & dist_trainer 2019-11-21 12:22:50 +08:00
yh_cc
1d7d05716e Merge branch 'master' of github.com:fastnlp/fastNLP 2019-11-21 11:31:28 +08:00
yh_cc
38fea0fcfa 1.增强CMRC2018Metric的使用范围; 2. StaticEmbedding支持仅使用来自于pretrain embedding的token 2019-11-21 11:31:21 +08:00
Yige Xu
81c71aceb8 [bugfix]修复了trainer在update_every大于1时loss的显示错误 2019-11-17 01:27:54 +08:00
Yige Xu
3510700d73 [bugfix]修复了一些文档错误 2019-11-15 23:21:52 +08:00
yunfan
a3738b3d3c [bugfix] 针对pytorch1.3.0版本bug的补丁 2019-11-06 13:39:44 +08:00
yhcc
fb1ad7a8fa
Merge pull request #246 from fastnlp/dev0.5.0 
修复针对pytorch1.3.0版本bug的一个bug
 2019-11-06 11:05:13 +08:00
Yige Xu
6cb3713b2c 修复针对pytorch1.3.0版本bug的一个bug 2019-11-06 01:26:43 +08:00
yhcc
3761c32beb
修复寻找hyper-parameter数量的bug 2019-10-30 17:06:56 +08:00
yhcc
4ceca47352
Merge pull request #245 from fastnlp/dev0.5.0 
[update]小幅度更新
 2019-10-25 22:28:30 +08:00