1ed1b314
Merge branch 'dev0.5.0' of https://github.com/fastnlp/fastNLP into dev0.5.0 by
2019-09-16 15:45:14 +0800
30180b91
Revert "delete online Data Getter and Iter" by
2019-09-16 15:45:02 +0800
d582bd3e
Delete train.tsv by
2019-09-16 15:40:18 +0800
f2face3b
Delete test.tsv by
2019-09-16 15:39:58 +0800
7b08c777
Delete dev.tsv by
2019-09-16 15:39:43 +0800
9b21071c
delete online Data Getter and Iter by
2019-09-16 15:36:59 +0800
bf2920cb
update reproduction/matching/matching_bert.py by
2019-09-16 15:09:20 +0800
6d8b4483
add tutorial by
2019-09-16 14:05:23 +0800
8614b89a
增加conll的pipe对bioes的检测 by
2019-09-15 22:34:22 +0800
e314a367
fix a bug function _indexize in fastNLP/io/pipe/utils.py by
2019-09-12 03:10:17 +0800
64fc8bc1
1. update classification and matching loader and pipe; 2. add data and test codes for testing classification and matching loader and pipe. by
2019-09-12 02:59:45 +0800
38d5b36b
Update test_callbacks.py by
2019-09-11 17:41:46 +0800
33cbb5b5
Create bert_embedding_tutorial.ipynb by
2019-09-11 16:51:19 +0800
7726c169
Update tester.py by
2019-09-11 16:48:55 +0800
a3b4d5e7
remove tensorboard logs after testing TensorboardCallback by
2019-09-11 14:10:01 +0800
f2299169
add assert in test_elmo_embedding by
2019-09-11 02:19:27 +0800
753327d2
fix code style in coreference task and related codes by
2019-09-11 02:16:57 +0800
143bf8ed
update DataBundle and add two property: num_dataset and num_vocab by
2019-09-11 02:14:12 +0800
5e6e35db
update README.md by
2019-09-11 02:13:12 +0800
9214cc6e
add the auto baseclass doc by
2019-09-10 23:41:12 +0800
202bde4b
split the docs over~ by
2019-09-10 22:56:50 +0800
32e99e36
split the docs in models by
2019-09-10 22:22:33 +0800
4a9cd850
split the docs in io subpackage by
2019-09-10 22:01:39 +0800
45fbbac7
split the docs in embeddings by
2019-09-10 19:11:18 +0800
e6f47819
Merge pull request #221 from Xiaoxiong-Liu/dev0.5.0 by
2019-09-10 18:31:37 +0800
b015cc14
undocumented by
2019-09-10 16:14:18 +0800
d481c84a
delete a out-date file by
2019-09-10 15:53:41 +0800
df4c5a3d
fix a lot of doc bugs in codes by
2019-09-10 15:52:03 +0800
9e3251df
fix a lot of doc bugs in codes by
2019-09-10 14:47:14 +0800
c9883bcb
undocumented by
2019-09-10 15:06:23 +0800
07bbb79f
Merge pull request #4 from fastnlp/dev0.5.0 by
2019-09-10 13:32:29 +0800
c38d815c
修改注释 by
2019-09-10 13:17:45 +0800
429dfe53
Merge branch 'dev0.5.0' of https://github.com/fastnlp/fastNLP into dev0.5.0 by
2019-09-10 12:48:31 +0800
9b8265dc
修复char_embedding bug by
2019-09-10 12:39:57 +0800
587edd54
update the doc-checking tool by
2019-09-10 11:08:27 +0800
164f74e1
Merge branch 'dev0.5.0' of https://github.com/fastnlp/fastNLP into dev0.5.0 by
2019-09-10 10:38:07 +0800
917cedf8
增加新的tutorial; 删除各embedding中requires_grad的设置,统一到基类实现 by
2019-09-10 10:38:02 +0800
0e0ba10a
Merge pull request #219 from HrWang96/loader-pr by
2019-09-09 22:18:18 +0800
9d31ded8
Merge branch 'dev0.5.0' into loader-pr by
2019-09-09 22:12:44 +0800
2fc14790
Merge pull request #217 from ZikaiGuo/dev0.5.0 by
2019-09-09 13:02:05 +0800
1caa83d0
Update transformer.py by
2019-09-08 14:54:31 +0200
77684043
loader-pr by
2019-09-08 16:28:53 +0800
53bcc0b2
loader-pr by
2019-09-08 16:28:15 +0800
8c8e22cc
DataSet中增加print_field_meta方法,使得其可以获取field的input和target信息 by
2019-09-07 18:47:03 +0800
04a54df2
Update requirements.txt by
2019-09-07 16:06:35 +0800
2fbc1d78
change the print format for dataset and instance by
2019-09-07 15:22:43 +0800
5bbfb92a
代码规范以及修改测试文件路径以匹配github文件路径 by
2019-09-06 13:08:57 +0800
f004a070
update the doc tool by
2019-09-05 15:13:08 +0800
5b7e9b65
update the ChnSentiCorpPipe in docs by
2019-09-05 14:32:37 +0800
b5a7db0b
delete the output part in dot-utils by
2019-09-05 14:31:38 +0800
ea5fbc88
增加注释 增加测试文件及测试样例 修改部分变量命名 by
2019-09-05 05:07:52 +0800
880e3ad9
1. add mini_elmo.pkl and test codes for testing ElmoEmbedding; 2. update bert testing codes by
2019-09-05 01:26:22 +0800
14d048f3
修复bert embedding的bug by
2019-09-05 00:21:46 +0800
60a535db
fix a little error in doc. by
2019-09-04 17:15:21 +0800
b1fe5f53
split the class's doc & __init__'s doc (core part) by
2019-09-04 16:53:31 +0800
a2e31584
update the auto alias tool by
2019-09-04 15:50:01 +0800
4caacade
delete the alias in files. by
2019-09-04 14:51:50 +0800
3651d61f
delete the alias in files. by
2019-09-04 14:47:45 +0800
113ef8b1
add code to detect the defined location automatically by
2019-09-04 14:31:45 +0800
e903db0e
增加中文分类Pipe;使用矩阵加速BertEmbedding部分pool_method;调整部分测试用例名称;修复metric中对warning的误报 by
2019-09-04 12:47:52 +0800
d15ad75d
fix a bug in test code by
2019-09-03 23:33:10 +0800
b3718b10
1. rename init_embed to embed in models/*; 2. update documents in models/bert.py; 3. update tutorial six. by
2019-09-03 23:19:18 +0800
8dae71ff
pipeline by
2019-09-03 14:19:29 +0800
53f744a8
fix some bugs in docs by
2019-09-02 19:43:28 +0800
1994029a
1.当前支持的encoding_type都支持从tag_vocab中自动判断;避免触发无意识导致的metric bug; 2. 修复部分inplace操作无法求导的问题; 3.Vocabulary将一些属性通过property暴露 by
2019-09-02 15:59:45 +0800
b9aa05f6
add testing codes and data for loader and pipe. by
2019-09-01 11:22:42 +0800
1c2ee50c
[fix] EchoCallback by
2019-09-01 10:31:14 +0800
091f24e3
fix some bugs in test code. by
2019-09-01 10:15:11 +0800
dc2885c6
修复sequence labeling 测试报错 by
2019-09-01 02:00:03 +0800
4440801d
1. update bert.py and fix a bug in bert_embedding to adapt torch 1.2.0; 2. update models/bert.py and add BertForSentenceMatching model, now a BertEmbedding param should be passed to these five models; 3. create a small bert version for testing and modify test/models/test_bert.py; 4. move small glove and word2vec files to data_for_tests/embedding/small_static_embedding dir and fix relevant test codes; 5. delete some __init__.py files in test dir. by
2019-09-01 01:19:10 +0800
bbda73c1
[update] transformer by
2019-08-30 21:48:00 +0800
44af6478
[update] change data-loader to pipe by
2019-08-29 20:19:13 +0800
82b57266
update transformer by
2019-08-24 13:59:30 +0800
9529f89a
增加DataBundle的方法;增加BilSTMCRF的注释 by
2019-08-30 19:54:28 +0800
b134c9f7
Merge pull request #211 from lyhuang18/lyhuang-reproduction by
2019-08-30 01:23:41 +0800
9e6f4ffb
datasetloader改成pipe by
2019-08-30 01:21:59 +0800
f7d6ea20
Merge pull request #18 from fastnlp/dev0.5.0 by
2019-08-30 00:48:47 +0800
0908c736
fix code in BertModel.from_pretrained and BertEmbedding by
2019-08-29 16:38:17 +0800
6f337c3f
Motified to Fastnlp/io/pipe/summarization.py by
2019-08-29 15:42:51 +0800
95d9a767
remove datapath by
2019-08-29 15:31:13 +0800
c29aca77
[verify]charcnn use pipe,remove dataloader by
2019-08-29 15:28:50 +0800
1756e3ff
1.修复MNLILoader中的bug; 2.修复field中的tensor warning by
2019-08-29 11:16:59 +0800
146a004d
修改travis converage设置 by
2019-08-29 10:12:30 +0800
09d0b745
Update .travis.yml by
2019-08-29 09:56:36 +0800
39de27f4
Update BertModel.from_pretrained function. Now can pass a model_dir_or_name instead of model_dir. by
2019-08-29 08:19:36 +0800
8142bad8
Merge branch 'dev0.5.0' of https://github.com/fastnlp/fastNLP into dev0.5.0 by
2019-08-29 01:43:19 +0800
5d8a8c98
1. delete io/data_loader dir; 2. delete model/enas*; 3. delete legacy dir; 4. delete DateSetLoader and relevant codes; 5. fix a test code error in core/test_dataset.py; 6. delete io.BaseLoader and relevant code. by
2019-08-29 01:43:04 +0800
cbe5b347
SpanFMetric增加对encoding_type和tag_vocab的检查 by
2019-08-28 23:53:53 +0800
55e736bf
SpanFMetric增加对encoding_type和tag_vocab的检查 by
2019-08-28 23:53:20 +0800
a46b8f12
Trainer中保存最佳模型存在bug by
2019-08-28 23:06:13 +0800
6201f661
Trainer中保存最佳模型存在bug by
2019-08-28 22:56:02 +0800
158721dc
change dataloader to pipe by
2019-08-27 21:48:09 +0800
fbbb2fcd
fix some bugs in docs by
2019-08-27 21:33:18 +0800
169f519f
ignore the methods inherited from torch.nn.Embedding by
2019-08-27 21:07:22 +0800
04737a10
update the doc-tool to show __init__ and class doc separately by
2019-08-27 20:46:05 +0800
359a1767
merge conflict by
2019-08-27 01:57:33 +0800
19bbaf11
使用更pytorch的方式处理embedding中的parameter by
2019-08-27 01:54:15 +0800
b4e54209
pipe by
2019-08-26 19:21:35 +0800
b5d04fd7
Merge pull request #3 from fastnlp/dev0.5.0 by
2019-08-26 17:50:08 +0800
53975c04
update the doc-tool & fix an importing bug by
2019-08-26 14:58:36 +0800