Skip to content

O
oneflow

Oneflow-Inc / oneflow
上一次同步 2 年多

通知 13
Star 2733
Fork 0
O
oneflow

前往新版Gitcode，体验更适合开发者的 AI 搜索 >>

切换分支/标签
  1. 14 11月, 2017 1 次提交
    • W
      reconstruct (#384) · d4694bbe
      由 Will Zhang 提交于 
      * branch init

* logical graph

* chain graph with bp struct

* backward done

* loss record struct

* build model struct

* compute task node

* fix bugf

* GenSortedCompTaskNodes

* clone op with boxing thread

* BuildSubTaskGraphMethod

* remove addclone for data_loader

* copy task node

* task graph build done

* fix boxing

* add TODO

* update

* update

* MaxModelSplitNum

* data part num

* DataConcatAndDataSplit


Former-commit-id: 8ed98d5c
      d4694bbe
  3. 19 10月, 2017 1 次提交
  5. 18 10月, 2017 3 次提交
    • duduscript's avatar
      Dev yongxian mdupdate (#367) · 5e04c1db
      由 duduscript 提交于 
      * fix label bug

* max_data_id_length

* no problem with dataloader kernel

* remove modify

* CHECK max_data_id

* data_id

* fix train problem

* fix some problem for data id

* fix

* Unified naming

* clang-format

* predict_job.proto

* concatKernel add data_id

* change \n position and fix some problem in kernel

* seperate copy data and copy dataid

* add paras in concat kernel

* solve record output problem

* remove changes in train_placement

* rename CopyDataIdFromIbToAllOb API and dataloader mathod

* unified naming

* unify naming

* unify naming

* unify naming

* change CopyDataIdFromSoleIbToAllObIfNeed API

* Fix CHECK range bug in concat.cpp

* set model update default value

* remove modify in example


Former-commit-id: 6903f359
      5e04c1db
    • W
      fix bug : append to record · 9ff19d5d
      由 willzhang4a58 提交于 
      
Former-commit-id: c24059bd
      9ff19d5d
    • duduscript's avatar
      Data Id (#357) · 26614ed1
      由 duduscript 提交于 
      * fix label bug

* max_data_id_length

* no problem with dataloader kernel

* remove modify

* CHECK max_data_id

* data_id

* fix train problem

* fix some problem for data id

* Unified naming

* clang-format

* predict_job.proto

* concatKernel add data_id

* change \n position and fix some problem in kernel

* seperate copy data and copy dataid

* add paras in concat kernel

* solve record output problem

* remove changes in train_placement

* rename CopyDataIdFromIbToAllOb API and dataloader mathod

* unified naming

* unify naming

* unify naming

* unify naming

* change CopyDataIdFromSoleIbToAllObIfNeed API

* Fix CHECK range bug in concat.cpp


Former-commit-id: 6db713e5
      26614ed1
  7. 17 10月, 2017 5 次提交
  9. 16 10月, 2017 2 次提交
  11. 15 10月, 2017 3 次提交
  13. 14 10月, 2017 1 次提交
  15. 12 10月, 2017 2 次提交
  17. 10 10月, 2017 4 次提交
  19. 09 10月, 2017 1 次提交
  21. 07 10月, 2017 1 次提交
  23. 30 9月, 2017 3 次提交
  25. 29 9月, 2017 4 次提交
  27. 28 9月, 2017 7 次提交
  29. 27 9月, 2017 2 次提交