hummingbird
/
fastNLP
Not watched
2
0
Fork 0
Code
Releases 13 Wiki Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain

Commit Graph

  • *
  • | *
  • | *
  • | *
  • | *
  • |/|
  • * |
  • | *
  • |/
  • *
  • *
  • |\
  • * |
  • | *
  • |/
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • |\
  • * |
  • | *
  • |/
  • *
  • *
  • *
  • |\
  • * \
  • |\ \
  • | | *
  • | |/
  • |/|
  • * |
  • * |
  • |\ \
  • * | |
  • | * |
  • | * |
  • | * |
  • | | *
  • * | |
  • |/ /
  • * |
  • | *
  • * |
  • * |
  • |\ \
  • * | |
  • | * |
  • |/ /
  • * |
  • * |
  • * |
  • |\ \
  • * | |
  • | * |
  • | * |
  • | * |
  • | * |
  • |/ /
  • * |
  • |\ \
  • * | |
  • | * |
  • |/ /
  • * |
  • * |
  • |\ \
  • * \ \
  • |\ \ \
  • | | * |
  • | * | |
  • | * | |
  • * | | |
  • |/ / /
  • * | |
  • | * |
  • |/ /
  • * |
  • * |
  • * |
  • * |
  • * |
  • |\ \
  • | * |
  • | * |
  • | * |
  • |/ /
  • * |
  • * |
  • * |
  • * |
  • * |
  • * |
  • * |
  • |\ \
  • * | |
  • | * |
  • |/ /
  • * |
  • * |
  • * |
  • |\ \
  • * | |
  • | * |
  • | * |
  • | |\ \
  • | * | |
  • | | | *
  • | | * |
  • | |/ /
  • | | *
  • | | *
  • | * |
  • | * |
  • | * |
  • | * |
  • | * |
  • | * |
  • |/ /
  • * |
  • * |
  • * |
  • * |
  • |\ \
  • * | |
  • | * |
  • | |\ \
  • | | * |
  • 9a9732a8 Update vocabulary.py (#325) by Gosicfly 2020-10-11 10:59:52 +0800
  • 0ca22b6b 1.修改部分文档中的typo; 2.支持对transformers embedding by yh_cc 2020-10-25 13:56:54 +0800
  • 81bcb513 当activation=lambda x: x出现错误 (#330) by w.g 2020-10-22 09:41:22 +0800
  • 3270b8b4 token_type_id_rev (#329) by stratoes 2020-10-20 00:20:32 +0800
  • 08cf1a51 Merge branch 'master' of github.com:fastnlp/fastNLP by yh_cc 2020-10-12 13:57:08 +0800
  • ba230455 update travis by yh_cc 2020-10-12 13:56:58 +0800
  • 0c990e7b Update vocabulary.py (#325) by Gosicfly 2020-10-11 10:59:52 +0800
  • edd3022b 新增ConstantTokenNumSampler使得每个sample中token数量一致，可以最大化利用GPU的显存 by yh_cc 2020-09-28 10:21:39 +0800
  • fe7ce8e4 Merge branch 'master' of github.com:fastnlp/fastNLP by yh_cc 2020-09-21 20:17:00 +0800
  • ee584e71 修复seq2seq_generator的bug by yh_cc 2020-09-21 20:16:43 +0800
  • 8d08626d [update] fix padding masking in viterbi_decode() of CRF by yunfan 2020-09-21 13:14:26 +0800
  • 0d0a6f74 修复测试失败问题 by yh_cc 2020-09-19 14:38:26 +0800
  • b6a07be3 [test] fix f1 test (#314) by ROGERDJQ 2020-09-19 11:36:07 +0800
  • 40bec216 修复save_path在dev为空的时候的bug by yh_cc 2020-09-19 11:10:36 +0800
  • 720d7b87 Update .travis.yml by Yining Zheng 2020-09-13 20:29:52 +0800
  • acdebfcc [bugfix]修复 coreference resolution复现代码中参数名字不对应的bug (#323) by liuxiaoxiong 2020-08-29 16:16:57 +0800
  • b9b688d2 f1 fix by roger 2020-07-30 04:24:28 +0000
  • 14ea1256 Merge branch 'master' of github.com:fastnlp/fastNLP by yh_cc 2020-07-29 15:02:12 +0800
  • 206f7758 1.新增一个sortedSampler，可以用于测试的时候加速; 2.修改BeamSearch中的bug; 3.修改Trainer中的循环代码，解决BatchSampler场景下可能出现evaluate时机不对的问题 by yh_cc 2020-07-29 15:02:05 +0800
  • 228cca44 [new] add show-result-list for confusion_matrix (#309) by ROGERDJQ 2020-07-10 14:28:30 +0800
  • c957ed69 loader中load先将filename排序保证一致的load顺序, 避免测试不通过 by yh_cc 2020-07-07 22:53:05 +0800
  • ed230971 修复一个data_parallel情况下的gpu不从0开始导致的bug by yh_cc 2020-07-05 23:24:34 +0800
  • a6c2533b Merge branch 'master' of github.com:fastnlp/fastNLP by yh_cc 2020-06-28 14:38:11 +0800
  • e361b32c 在@linzehui 的帮助下seq2seq终于有了第一个版本; 目前实现了Seq2Seq的Transformer和LSTM版本，但metric和loss还没update; by yh_cc 2020-06-28 14:38:02 +0800
  • d9456eb1 [bugfix] fix an import error in bert_tokenizer.py by Yige Xu 2020-06-18 22:51:52 +0800
  • 94be3315 [bugfix]修改了modules/utils.py中`initial_parameter()`函数的bug (#300) by 郑恒毅 2020-06-07 22:24:36 +0800
  • 1efd40c0 准备开发0.5.6 by yh_cc 2020-06-07 20:37:25 +0800
  • 43300ad9 修改部分注释, 准备开发0.5.6 by yh_cc 2020-06-07 20:35:09 +0800
  • 8732dfd9 新增了 DataBundle 的 apply_more 和 apply_field_more 方法。需要进一步试用和测试。 by ChenXin 2020-06-06 11:06:17 +0800
  • 1f27d007 [add] add MSELoss by Yige Xu 2020-05-21 12:48:08 +0800
  • 4e95989e [new] add seperator for conll loader (#293) by ROGERDJQ 2020-05-03 19:46:11 +0800
  • b95aa56a 基本完成seq2seq基础功能 by linzehui 2020-05-03 15:52:10 +0800
  • 6c7a61a9 (tag: v0.5.5) 发布0.5.5版本 by yh_cc 2020-05-02 11:57:25 +0800
  • ae7b9163 1.修改Generator中的采样，防止导致分布全0采样失败; 2. 修改Trainer与checkpointcallback的配合 by yh_cc 2020-05-02 11:43:34 +0800
  • 15360e97 保存一版旧版 by linzehui 2020-04-30 17:27:49 +0800
  • 9b45e9c0 修改GradientClipCallback中parameter的存储方式，防止在torch 1.5版本中报错 by yh_cc 2020-04-29 00:14:34 +0800
  • 3683fd5b Merge branch 'master' of github.com:fastnlp/fastNLP by yh_cc 2020-04-29 00:13:31 +0800
  • 62fe53b1 修改GradientClipCallback中parameter的存储方式，防止在torch 1.5版本中报错 by yh_cc 2020-04-29 00:13:13 +0800
  • 1a1254b2 修改roberta embedding的一个文档错误 by Yige Xu 2020-04-28 23:35:12 +0800
  • fbd2fd4e update DataSet使得可以与之前老版本的DataSet兼容 by yh_cc 2020-04-26 12:32:46 +0800
  • ff8b9a37 修复Embedding注释错误与core.utils.py代码错误 by yh_cc 2020-04-18 16:12:11 +0800
  • 91d9b239 Merge branch 'master' of github.com:fastnlp/fastNLP by yh_cc 2020-04-14 15:35:13 +0800
  • 4dd16be8 修复vocabulary iterate时不符合常规情况的问题 by yh_cc 2020-04-14 15:35:05 +0800
  • 45572020 更新了 ControlC 的 API by ChenXin 2020-04-14 10:51:23 +0800
  • 7cdd2ee8 在文档中增加了分布式训练的介绍（by yunfan） by ChenXin 2020-04-14 09:23:59 +0800
  • a1ac3562 [update] dist_trainer tutorial by yunfan 2020-04-13 23:25:44 +0800
  • d997e5c7 [bugfix] fix start_time, save&load in dist_trainer by yunfan 2020-04-13 23:22:06 +0800
  • d6b9c59c Merge branch 'master' of github.com:fastnlp/fastNLP by yh_cc 2020-04-13 17:04:10 +0800
  • 18747e63 修复Vocabulary在load的时候可能发生的bug by yh_cc 2020-04-13 17:03:53 +0800
  • f0e1bdad [bugfix] fix an import error by Yige Xu 2020-04-12 22:26:27 +0800
  • 2dee6712 Solve the GPTTokenizer dependency problem by yh_cc 2020-04-12 21:13:33 +0800
  • ab047a2a 新增对RobertaEmbedding和GPT2Embedding的支持 by yh_cc 2020-04-12 00:49:02 +0800
  • cbbfc181 Merge branch 'master' of github.com:fastnlp/fastNLP by yh_cc 2020-04-11 22:56:09 +0800
  • 7398f2b0 1.增加RobertaEmbedding与GPT2Embedding by yh_cc 2020-04-11 22:55:54 +0800
  • 8a084c4f [update] bugfix in dist_trainer by yunfan 2020-04-11 17:01:17 +0800
  • b49c694c [update] bugfix in dist_trainer by yunfan 2020-04-11 16:20:42 +0800
  • 45139ebb BertEmbedding支持选择是否缩小词表 by yh_cc 2020-04-07 01:07:52 +0800
  • ace8aafd 修复BertEmbedding在include_cls_sep=True时，SEP位置的值不正确的问题 by yh_cc 2020-04-06 01:40:26 +0800
  • d2ca4663 upload init roberta embedding by Yige Xu 2020-04-05 21:50:44 +0800
  • ba645f44 1.将collect_fn修改为collate_fn使得与pytorch保持一致; 2.增加对BatchSampler的支持 by yh_cc 2020-04-04 14:52:03 +0800
  • 9b77ad81 暴露更多类 by yh_cc 2020-03-28 21:41:53 +0800
  • 3b0e5e06 暴露更多内容 by yh_cc 2020-03-28 21:40:32 +0800
  • eba9121a 对外暴露LossBase by yh_cc 2020-03-28 21:36:34 +0800
  • 4860442a Merge pull request #279 from fastnlp/document-fix by ChenXin 2020-03-27 22:01:46 +0800
  • 693981be 修改了文档的其它bug by ChenXin 2020-03-27 21:29:58 +0800
  • 36d68e92 修改.py文件中的文档格式 by ChenXin 2020-03-27 20:45:04 +0800
  • c371754b use r""" ... """ replace """ ... """ by ChenXin 2020-03-27 20:43:45 +0800
  • a096e702 修复了一系列文档中的 bug by ChenXin 2020-03-27 01:46:20 +0800
  • bb018369 使代码文件可以自动下载 by ChenXin 2020-03-27 00:28:15 +0800
  • 347efced 增加了教程的代码下载 by ChenXin 2020-03-26 23:01:18 +0800
  • 367edb7a 修复move_to_device中device为None时可能出现错误的问题; BERTEmbedding中增加一个可选参数 by yh_cc 2020-03-26 20:04:03 +0800
  • a8aa1035 修复batch转tensor的bug; 新增对TorchLoaderIter的说明 by yh_cc 2020-03-25 12:33:47 +0800
  • 885c7402 增加对ConfusionMatrix的测试用例 by yh_cc 2020-03-24 16:09:00 +0800
  • b3ace23d Merge branch 'master' of github.com:fastnlp/fastNLP by yh_cc 2020-03-24 15:53:41 +0800
  • bbb1dbb6 修改tutorial中的typo; 增加DataBundle的add_collect_fn by yh_cc 2020-03-24 15:52:41 +0800
  • 13428adf [new]revised confusion class (#278) by ROGERDJQ 2020-03-24 15:51:53 +0800
  • bb88bb4f convert numpy to tensor by yh_cc 2020-03-19 18:49:41 +0800
  • 7f019713 调整涉及到多个field取batch的实现方式 by yh_cc 2020-03-19 18:04:13 +0800
  • dce1a73f Merge branch 'master' of github.com:fastnlp/fastNLP by yh_cc 2020-03-19 14:22:22 +0800
  • 8fe7a4f1 防止BERTEmbedding在中文场景下被错误使用 by yh_cc 2020-03-19 14:22:15 +0800
  • 75d59370 [update] multi criteria cws by yunfan 2020-03-18 11:26:00 +0800
  • f65843d7 Merge branch 'master' of https://github.com/fastnlp/fastNLP by Yige Xu 2020-03-16 19:30:42 +0800
  • 56bf05ad Update test_classification.py by Yige Xu 2020-03-16 19:29:36 +0800
  • 7000cdeb add t2t model by linzehui 2020-03-16 19:26:08 +0800
  • 9577a9ac fix a tiny bug by ChenXin 2020-03-16 18:57:21 +0800
  • 879cef62 add lstm module and generator by linzehui 2020-03-16 18:20:08 +0800
  • 8ae58d15 init commit, finish basic T2T component by linzehui 2020-03-16 15:56:58 +0800
  • 918a4137 使用 callable 替代 isfunction by ChenXin 2020-03-16 11:50:15 +0800
  • d2cc93cd 使用 callable 替代 isfunction by ChenXin 2020-03-16 11:34:42 +0800
  • 26ee41f9 DataSet 增加了 apply_more 和 apply_filed_more 两个新方法 by ChenXin 2020-03-16 10:42:56 +0800
  • 54dcc23a [update] collect_fn keep tracks for field by yunfan 2020-03-15 15:46:45 +0800
  • 2b41e4dd [bugfix] auto convert tensor type when batching by yunfan 2020-03-15 15:19:42 +0800
  • 65b141c1 Update bert.py by Yige Xu 2020-03-14 21:03:00 +0800
  • f35a4ae2 在没有编写模块文档的情况下不会报错，而是改为提示。 by ChenXin 2020-03-13 23:36:24 +0800
  • 423e8e37 [update] huge modify on collect_fn by yunfan 2020-03-13 21:23:18 +0800
  • 127d6d19 [add] collect_fn for multi field combined inputs by yunfan 2020-03-07 23:16:15 +0800
  • 5ffd46ae Merge branch 'master' of github.com:fastnlp/fastNLP by yh_cc 2020-03-06 23:43:17 +0800
  • 8acb0f4a 1.新增Vocabulary的save与load功能; 2.修复BERTEmbedding drop_word的bug; 3.新增各种Embedding的使用注释 by yh_cc 2020-03-06 23:43:08 +0800
  • 10204ef0 Merge pull request #273 from fastnlp/tutorial-fix by ChenXin 2020-03-04 23:21:08 +0800
  • cb2aecf7 更新了 tutorial 的 README，准备合入 master 分支 by ChenXin 2020-03-04 23:07:11 +0800