hummingbird
/
fastNLP
Not watched
2
0
Fork 0
Code
Releases 13 Wiki Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain
2431 Commits
3 Branches
25 MB
139 Download
Tree: c11c30a5fc
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'c11c30a5fc'
${ noResults }
Commit Graph

332 Commits (c11c30a5fc8366c04f4fa445bafcc675c1b2ad79)

Author SHA1 Message Date
  x54-729 5a41cca483 seq_len_to_mask 添加oneflow部分；完善oneflow的测试标签 3 years ago
  MorningForest 12a3c45344 Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 3 years ago
  MorningForest a3adafea34 add dataset method 3 years ago
  YWMditto 4b6e455247 更新了文档 3 years ago
  YWMditto b1e23de8f7 Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 3 years ago
  YWMditto 0506fc2fcb 添加了 TorchFSDPDriver；修改了 ddp 中的部分细节；删除了 topksaveer 的 rank_zero_only 修饰器 3 years ago
  x54-729 5e7941f6a8 Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 3 years ago
  x54-729 79b42a91ce 调整 fastNLP/core/drivers 的文档 3 years ago
  MorningForest 0098c0c896 pipe增加num_proc 3 years ago
  x54-729 f5005e1e8b fix conflict 3 years ago
  x54-729 eb218f3def 调整 driver 的部分文档 3 years ago
  x54-729 4cdfdc77c7 oneflow 动态图 单卡和ddp部分及测试 3 years ago
  x54-729 ad6ada2487 small 3 years ago
  x54-729 a495bb938a 1.修复模型会被移动到rank对应设备的问题 2.更改 deepspeed driver æ˜命名 3.为 deepspeed 添加 logging_level 3 years ago
  x54-729 2735d2d10c DeepSpeedDriver现在可以通过 deepspeed 命令拉起；添加了相关 trainer 的简单测试 3 years ago
  x54-729 7023ea550c deepspeed的save load功能 3 years ago
  x54-729 22d95be007 Merge branch 'dev0.8.0' into deepspeed 3 years ago
  x54-729 b60621f3d1 small 3 years ago
  x54-729 cbee1c6cbc deepspeed test init 3 years ago
  x54-729 a39d3011e8 Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 3 years ago
  YWMditto bb68856f85 添加对 overfit 多卡的测试 3 years ago
  YWMditto 024fecfbf3 添加了 overfit 的功能 3 years ago
  x54-729 cd0957fb5b 跟进paddle jittor 关于 set_dist_repro_dataloader函数中的修改 3 years ago
  yhcc dd1c5ca035 修复设置了global seed的bug 3 years ago
  yhcc 7283bf27b2 在私有定制Sampler的情况下，多卡不替换 3 years ago
  x54-729 c99315f79e 修复新增的set_dist_repro_dataloader函数测试例在paddle情况下的问题 3 years ago
  yhcc 47b7c4e832 增加set_dist_repro_dataloader测试 3 years ago
  x54-729 4cc8a29926 1.为element测试添加torch标签 2.解决torch版本导致的int张量无法对整数使用除法的问题 3 years ago
  yhcc 24092e3114 增加pipe相关的测试 3 years ago
  MorningForest ec17562f20 修改fdl 的prepare_*_dataloader,使其支持fastnlp的dataset 3 years ago
  MorningForest af6300e816 修改element,计算值转为非tensor形式 3 years ago
  yhcc 057806de7d 将total_batches改名为n_batches，以和n_epochs对应;增加n_batches作为Trainer的初始化参数 3 years ago
  yhcc 64c7ce5468 1.修复rich在jupyter的时候不打印的问题;2.修复若干其他bug 3 years ago
  x54-729 809b83b6c0 small 3 years ago
  x54-729 f1cb6f6167 small 3 years ago
  x54-729 edda3d3196 1.处理新测试例的import问题 2.将warnings.warn替换为logger.warn 3 years ago
  yhcc 6d889c8624 Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 3 years ago
  yhcc 004c344e4c 1.增加torch modules；torch models 3 years ago
  x54-729 9b0a30c8fb small bug 3 years ago
  yh 1d665c9480 unrepeatedSampler支持chunk_split 3 years ago
  x54-729 f7b3fe6a06 small 3 years ago
  x54-729 75a3278d69 1. 完成JittorSingleDriver的功能，并添加测试用例 2.在Sampler中添加属性num_samplers 用于动态获取dataset的长度 3.添加便于测试断点重训的数据集 4.修改jittor其它测试的一些bug，统一ArgMaxDataset 的命名 3 years ago
  x54-729 a6cfc4086f 修改seq_len_to_mask的jittor实现及测试例中的一处传参错误 3 years ago
  x54-729 999520e4b0 Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 3 years ago
  x54-729 28eb1a5836 添加了对tensor_to_numeric reduce 参数的测试 3 years ago
  yh 7a7d75d100 Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0 3 years ago
  yh dcdd484eb2 增加seq_len_to_mask 3 years ago
  Letian Li 005b0e055e
实现 jittor driver 多卡训练 (#418) 3 years ago
  Letian Li 5425095cac
[bugfix] 修复 tests/core/controllers/_test_trainer_jittor.py，使其可以正常运行 (#415) 3 years ago
  x54-729 da0b747b30 修改test_mixdataloader中的工具函数名，防止被pytest执行 3 years ago