hummingbird
/
fastNLP
Not watched
2
0
Fork 0
Code
Releases 13 Wiki Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain
2198 Commits
3 Branches
25 MB
866 Download
Tree: b1ba42e831
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'b1ba42e831'
${ noResults }
Commit Graph

255 Commits (b1ba42e8315e0dfbd0cb3440e011570ae3e8a801)

Author SHA1 Message Date
  yh b1ba42e831 修复若干微小的bug 4 years ago
  yh 6b56d8a082 merge callback test to speed up 4 years ago
  yh d4bccf3f6a 修复Evaluator的evaluate_use_dist_sampler在Trainer中的bug 4 years ago
  x54-729 c4af9f21c6 paddle 和测试例跟进 set_dist_repro_dataloader 函数；修改test_trainer_wo_evaluator_torch.py的bug 4 years ago
  x54-729 efaec85c57 1. 处理 topk_saver topk 参数为 None 的情况 2.Trainer on_exception 函数参数判断调整 3. 处理各类型 padder 中数据的 batch_size 为1 的情况 4. ReproducibleBatchSampler和ReproducibleSampler 对参数 length 的判断 5.TorchDataLoader 初始化参数调整 6.部分测试例跟进版本 4 years ago
  x54-729 6e66fb899e 修正 JittorDataLoader 读取 jittor Dataset 无法进行索引的问题 4 years ago
  yh 2423641010 新增TqdmProgressBar 4 years ago
  MorningForest 6488f5e00f 修改fdl冲突 4 years ago
  x54-729 655e48de99 1.修改torch测试例中多卡的driver参数 2.修改测试例中的driver.save driver.load为driver.save_checkpoint driver.load_checkpoint 3.添加lstm 4 years ago
  yh_cc 97b504c020 修改Trainer.save/Driver.save为Trainer.save_checkpoint/Driver.save_checkpoint 4 years ago
  MorningForest 2baedee74c Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 4 years ago
  yh_cc a470994686 print bug fix 4 years ago
  MorningForest 6340de06ea Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 4 years ago
  x54-729 3533bc2704 恢复 torch paddle jittor 之间的转换函数 4 years ago
  x54-729 d79de6b008 1. 继续完善部分文档；2.删除 paddle 多卡下的 data_device 功能 3. 将 paddle_utils 下的 get_device_from_visible 函数更名为 _convert_data_device 并进行修改 4 years ago
  x54-729 7f42451d6d Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 4 years ago
  yh_cc 7763b2e087 新增Callback on_load_checkpoint测试 4 years ago
  x54-729 ef892a7aed 1. 支持在不设置backend的情况下运行单卡的paddle程序 2.当通过launch启动且限制显卡时的paddle多卡逻辑 4 years ago
  MorningForest 6f96d86e9f Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 4 years ago
  x54-729 3af4732f55 Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 4 years ago
  yh_cc 5d8d577399 Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 4 years ago
  yh_cc cb5aa37015 补充部分文档 4 years ago
  x54-729 2a44af2519 为 test_trainer_jittor 添加 DummyClass 和 pytest.mark.jittor 4 years ago
  x54-729 34698f0f9e 删除 mix_module 和 torch_paddle_driver 的内容 4 years ago
  MorningForest 7c268fe546 Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 4 years ago
  x54-729 fe449bd543 删除 initialize_paddle_driver 对参数 driver的限制，使得它能够根据 device 初始化 driver 4 years ago
  MorningForest c90c489cd5 修改冲突 4 years ago
  YWMditto 3f00642ba1 修复了冲突 4 years ago
  YWMditto bcc66b43cf 修改了 trainer 和 evaluator 的 torch_kwargs 4 years ago
  yh_cc 47a63c27f7 1.修改部分注释; 2.增加prepare_dataloader函数;3.paddle的paddler修改为先集中使用numpy pad再一把to_tensor 4 years ago
  MorningForest 81ef1ec4b7 修改paddle fdl测试用例 4 years ago
  MorningForest 4c3b05b4b3 Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 4 years ago
  MorningForest af8fdf9080 修改fdl测试用例 4 years ago
  yh_cc f1fa665e7c merge 4 years ago
  yh_cc 22c4a6ef3d 1.ResultsMonitor;2.修改了Evaluator中metric名称的实现方式 4 years ago
  Letian Li a294955e32
[bugfix] 支持 Jittor single driver，并添加测试用例 (#413) 4 years ago
  MorningForest 9f6b9544a0 Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 4 years ago
  MorningForest 1ecf423013 修改测试用例 4 years ago
  yh_cc 2a80a9de44 Trainer.on支持*args等形式的输入 4 years ago
  yh_cc 47a4ab4d84 重命名tests/core/collators/test_pakcer_unpacker.py 4 years ago
  yh_cc 0fd17152b8 修改Collator的实现，优化了nested场景下的报错 4 years ago
  MorningForest cdf3474e2e 修改fdl测试用例 4 years ago
  MorningForest da25fbee3f 修改fdl 4 years ago
  MorningForest 97a7532a71 修改dl, dataset 4 years ago
  yh_cc 2297fcb30a Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 4 years ago
  yh_cc 8c5250f5b4 增强了cache_results， 现在可以将函数的参数hash进入保存的文件名字中 4 years ago
  x54-729 281a570b09 1、调整单卡中 save_and_load_model 测试例，不再使用 pytest.fixture 2、添加 PaddleFleetDriver 中 broadcast 误删的设备转换 4 years ago
  x54-729 38e578f6b5 PaddleDataLoader 1、初始化时如果 batch_sampler 为 None 时重置 shuffle 等参数 2、测试例中删除了不必要的 test_v2 4 years ago
  x54-729 cbffa896c5 删除 test_get_padder_run 中 jittor 的情况 4 years ago
  x54-729 45d08ee218 fix conflict 4 years ago