hummingbird
/
fastNLP
Not watched
2
0
Fork 0
Code
Releases 13 Wiki Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain
1684 Commits
3 Branches
25 MB
690 Download
Tree: 0ca22b6b17
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from '0ca22b6b17'
${ noResults }
Commit Graph

405 Commits (0ca22b6b17a6bf84811772b682a57b94de7a0968)

Author SHA1 Message Date
  yh_cc e903db0e70 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 6 years ago
  Yige Xu d15ad75d96 fix a bug in test code 6 years ago
  Yige Xu b3718b10dc 1. rename init_embed to embed in models/*; 2. update documents in models/bert.py; 3. update tutorial six. 6 years ago
  yh 1994029ab8 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 6 years ago
  Yige Xu b9aa05f6cf add testing codes and data for loader and pipe. 6 years ago
  Yige Xu 091f24e393 fix some bugs in test code. 6 years ago
  yh dc2885c6bc 修复sequence labeling 测试报错 6 years ago
  Yige Xu 4440801dbf 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. 6 years ago
  xuyige 8142bad87a Merge branch 'dev0.5.0' of https://github.com/fastnlp/fastNLP into dev0.5.0 6 years ago
  xuyige 5d8a8c98c6 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. 6 years ago
  yh_cc 55e736bf4c SpanFMetric增加对encoding_type和tag_vocab的检查 6 years ago
  yh_cc be77533c38 修复word drop bug, 增加相应测试 6 years ago
  xuyige 65a6fd3dc7 Revert "delete predictor.py" 6 years ago
  ChenXin 8445bdbc79 delete predictor.py 6 years ago
  yh_cc 85f01f01d1 1.修复部分测试; 2.修复StaticEmbedding中未找到词初始化bug 6 years ago
  yh e2232ac39f 1.修复Embedding中潜在的寻找路径失败; 2.reproduction中增加分词模型; 3.修改ner模型为最新的pipe版本 6 years ago
  yh 511f41dda1 1. 增加中文NER相关的loader和pipe; 2. 对应修改sequence_labeling的代码; 3.增加部分测试代码 6 years ago
  xuyige 9560a4d367 update test codes in models/bert.py 6 years ago
  ChenXin 1faf4ba2fa delete a out-date test case 6 years ago
  yh fd37ed60a7 1. Trainer增加一个dev_batch_size参数;2.StaticEmbedding中增加min_freq; 6 years ago
  yh c9fba2ae96 增加对static_embed的测试 6 years ago
  yh 3ae383efc3 修复BertEmbedding中长度会额外加长的bug 6 years ago
  yh 014e9786c7 1. 分类DataSetLoader中的Loader功能Pipe功能; 2. 增加数据集自动下载; 3.修复vocabulary中的bug 6 years ago
  yunfan cacf40366c [fix] distributed trainer 6 years ago
  yunfan 329a18976f [update] distributed trainer, add evaluation part 6 years ago
  yunfan 606d63a5a4 [update] distributed trainer 6 years ago
  yunfan 861f5387a4 [add] very first version of distributed trainer 6 years ago
  xuyige c2d687528e fix bugs and add test codes for: 1. models.snli; 2. core.metrics.extractive_qa; 3. io.data_loader.mnli 6 years ago
  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;“ 6 years ago
  yh 39f3accace 增加Embedding的文档 6 years ago
  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 6 years ago
  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 6 years ago
  xuyige 65f92855ba add testing code in stack embedding 6 years ago
  xuyige 327833a77f fix bug in test code 6 years ago
  xuyige 570b214dfb 增加fastNLP.embeddings模块并修改对应的现有代码以适配fastNLP.embeddings 6 years ago
  xuyige 2eba991660 fix matching DataLoader test code 6 years ago
  xuyige 7d07b38e0a fix bug in matching DataLoader 6 years ago
  yh_cc 1babf5384f Vocabulary中no_create_entry的bug修复 6 years ago
  xuyige a83cee0fbb fix bug in load dataset test code 6 years ago
  xuyige 66a7cf084e fix bug in test 6 years ago
  xuyige 089009f9f2 大幅度更新： 6 years ago
  yh_cc 84b18890a1 1.增加AdamW的optimizer;2.修复Trainer中metric_key的bug;3.静态embedding初始化修改;4.CrossEntropyLoss增加对reduction的支持' 6 years ago
  yh_cc e57b8e4fd3 seq_len_to_mask修复测试失败的问题 6 years ago
  yh_cc 8f7ed07441 1. 在vocabulary的from_dataset中增加no_create_entry_dataset选项，用于传递dev和test 6 years ago
  yh 8a766f070b seq_len_to_mask修改为直接使用max_len而不再和句中最大长度对比 6 years ago
  yh_cc 2f5d8967a3 1. 适配将Batch修改为pytorch的DataLoader的修改 6 years ago
  yh_cc 839d712467 增强field中的value_count支持对nested的field的支持 6 years ago
  yh_cc 17b5fd0066 1. 删除Trainer中对train_data必须为DataSet的assert 6 years ago
  yunfan efe3574014 Merge remote-tracking branch 'origin/dev0.5.0' into batch 6 years ago
  yunfan 7564818f4b [unstable] change Batch to torch's DataLoader 6 years ago