yh
|
af55db2019
|
1.修改callback prepare_callbacks, 2.让Dist_Trainer支持find_unused_parameters, 但仅在1.1以上版本有效
|
2019-07-29 23:56:53 +08:00 |
|
yh
|
71c9e0c30e
|
从远程下载权重
|
2019-07-29 23:54:57 +08:00 |
|
yh
|
db8c6a0b8a
|
解决读取bert权重潜在的bug问题
|
2019-07-26 19:44:20 +08:00 |
|
yh
|
d401bd2208
|
sst loader中增加一列raw_words
|
2019-07-25 16:56:54 +08:00 |
|
yh
|
cb3cb8bc5c
|
修复BertEmbedding的weight
|
2019-07-25 16:51:59 +08:00 |
|
yh
|
a8111087a3
|
删除getattr的default keyword
|
2019-07-25 16:36:28 +08:00 |
|
yh
|
210af73c6e
|
修复callback连续两次加入的bug
|
2019-07-25 16:01:17 +08:00 |
|
yunfan
|
329a18976f
|
[update] distributed trainer, add evaluation part
|
2019-07-20 17:12:15 +08:00 |
|
yunfan
|
606d63a5a4
|
[update] distributed trainer
|
2019-07-20 16:24:12 +08:00 |
|
yunfan
|
861f5387a4
|
[add] very first version of distributed trainer
|
2019-07-19 22:40:26 +08:00 |
|
yh
|
4718804e22
|
1. 修改内部函数位置; 2.修改BertWordPieceEncoder的部分功能
|
2019-07-19 17:33:41 +08:00 |
|
xuyige
|
c2d687528e
|
fix bugs and add test codes for: 1. models.snli; 2. core.metrics.extractive_qa; 3. io.data_loader.mnli
|
2019-07-19 02:05:33 +08:00 |
|
yh
|
22a8702d22
|
1. Trainer支持使用DistributedDataParallel训练; 但是还没有经过广泛测试,谨慎使用; 2. 修复import os bug; 3.FitlogCallback支持不传入任何DataSet; 4. NullOptimizer的construct_from_optimer返回self; 5. 修复Bert中pooled_cls的bug;“
|
2019-07-18 23:43:10 +08:00 |
|
yh
|
c19499e60a
|
1. 修复DataSet.delete_instance的bug; 2. FieldArray中支持只使用第一个instance推断dimension和type,节省时间
|
2019-07-18 22:53:30 +08:00 |
|
yh
|
3c2e419059
|
增加pooled_cls选项,可以是Bert在做分类时可以使用预训练的权重
|
2019-07-17 20:06:40 +08:00 |
|
ChenXin
|
d9e9f9ca7c
|
delete a todo item
|
2019-07-12 15:24:18 +08:00 |
|
xuyige
|
ec824376c6
|
fix embedding tutorials
|
2019-07-12 14:51:11 +08:00 |
|
Yige XU
|
25421d104d
|
Merge branch 'master' into dev0.5.0
|
2019-07-12 14:30:59 +08:00 |
|
ChenXin
|
1fed3650d5
|
check了 embeddings 的文档
|
2019-07-12 14:18:04 +08:00 |
|
xuyige
|
98ebc3c0b1
|
在README.md添加版本说明
|
2019-07-12 14:12:13 +08:00 |
|
yh
|
f3e19ddd9a
|
erge conflict in bert
|
2019-07-12 13:29:54 +08:00 |
|
yh
|
39f3accace
|
增加Embedding的文档
|
2019-07-12 13:27:29 +08:00 |
|
ChenXin
|
af4fd46273
|
修改了io 模块内子模块的顺序
|
2019-07-12 13:16:42 +08:00 |
|
ChenXin
|
90c6454aed
|
修改了文档首页
|
2019-07-12 13:13:54 +08:00 |
|
yunfan
|
9f681dcd55
|
fix tutorial typo
|
2019-07-12 12:42:31 +08:00 |
|
ChenXin
|
fc0f86aa9e
|
io、data_loader的文档结构和别名
|
2019-07-12 12:05:04 +08:00 |
|
ChenXin
|
ce72936c0f
|
:maxdepth: 1
|
2019-07-12 11:38:51 +08:00 |
|
ChenXin
|
dcc6d5d15d
|
models的文档结构和别名
|
2019-07-12 11:27:35 +08:00 |
|
ChenXin
|
f3a9fc5b79
|
encoder里面的结构和文档
|
2019-07-12 11:16:21 +08:00 |
|
ChenXin
|
a09cf518d7
|
modules入口的介绍和dropout的文档
|
2019-07-12 11:05:57 +08:00 |
|
ChenXin
|
d6ae241bbb
|
decoder部分的别名
|
2019-07-12 10:49:18 +08:00 |
|
ChenXin
|
c99f02aa3b
|
API文档入口的介绍
|
2019-07-12 10:46:29 +08:00 |
|
ChenXin
|
391793a961
|
最新的docs结构
|
2019-07-12 10:36:46 +08:00 |
|
yh
|
0a33a32081
|
Merge branch 'dev0.5.0' of https://github.com/fastnlp/fastNLP into dev0.5.0
# Conflicts:
# fastNLP/modules/encoder/embedding.py
# reproduction/seqence_labelling/ner/train_ontonote.py
# reproduction/text_classification/model/lstm.py
|
2019-07-12 09:56:35 +08:00 |
|
yh
|
807a8f1fb3
|
1. BucketSampler不需要自己传入batch_size了,由Trainer自动设置
2. sequence labeling中增加BiLSTMCRF
3. embedding中修复bug
4. 将之前版本的reproduction移动到legacy文件夹
5. sequence labeling中增加中文NER
6. text classification中loader增强,修改lstm模型,增加train_bert.py
|
2019-07-12 09:27:11 +08:00 |
|
lyhuang18
|
026925097d
|
Merge pull request #191 from lyhuang18/lyhuang-reproduction
修改代码以适配新fastNLP.embeddings模块
|
2019-07-12 06:50:10 +08:00 |
|
lyhuang18
|
ffbba0f895
|
修改代码以适配新embeddings模块
|
2019-07-12 06:47:04 +08:00 |
|
lyhuang18
|
0d487707d9
|
Merge pull request #17 from fastnlp/dev0.5.0
Dev0.5.0
|
2019-07-12 06:00:17 +08:00 |
|
xuyige
|
defcaae4bd
|
将test_embed_loader的almost equal范围稍稍调大一些,以防止偶然的测试不通过
|
2019-07-12 04:39:17 +08:00 |
|
xuyige
|
bedb792b3c
|
将test_embed_loader的almost equal范围稍稍调大一些,以防止偶然的测试不通过
|
2019-07-12 04:33:32 +08:00 |
|
lyhuang18
|
4df322eb2d
|
Merge pull request #14 from fastnlp/dev0.5.0
Dev0.5.0
|
2019-07-12 04:30:20 +08:00 |
|
xuyige
|
65f92855ba
|
add testing code in stack embedding
|
2019-07-12 04:24:08 +08:00 |
|
xuyige
|
327833a77f
|
fix bug in test code
|
2019-07-12 04:11:03 +08:00 |
|
xuyige
|
76198acc5c
|
update README.md: add description in fastNLP.embeddings
|
2019-07-12 04:09:27 +08:00 |
|
xuyige
|
570b214dfb
|
增加fastNLP.embeddings模块并修改对应的现有代码以适配fastNLP.embeddings
|
2019-07-12 04:07:47 +08:00 |
|
xuyige
|
6cf1a8562b
|
rename Batch to DataSetIter in enas_trainer
|
2019-07-12 03:19:43 +08:00 |
|
xuyige
|
273657048f
|
更改文档内容
|
2019-07-12 01:47:51 +08:00 |
|
xuyige
|
9f3968d97b
|
Merge remote-tracking branch 'origin/dev0.5.0' into dev0.5.0
|
2019-07-12 00:15:49 +08:00 |
|
xuyige
|
579bdb1356
|
更新DataSetLoader的文档以及对应教程
|
2019-07-12 00:15:15 +08:00 |
|
xuyige
|
2e523c6096
|
更新fastNLP框架图及流程图
|
2019-07-12 00:14:11 +08:00 |
|