aboutsummaryrefslogtreecommitdiff
path: root/nerv/layer/affine.lua
Commit message (Collapse)AuthorAge
* fix bug of momentum & update mse layerQi Liu2016-03-29
|
* major change: clearer param binding semantics; permit rebinding; enableDeterminant2016-03-10
| | | | resuming from previous training
* fixed some minor problemtxh182016-02-16
|
* improve parameter auto-detectionDeterminant2016-02-16
|
* move curand_gen out of the matrix layoutDeterminant2016-02-15
|
* try a basic mergeDeterminant2016-02-15
|
* changed auto-generating params, won not save in global_conf.paramtxh182015-11-25
|
* let affine supported multiple inputstxh182015-11-24
|
* small bug fixtxh182015-11-23
|\
| * correct the use of self.gconfDeterminant2015-11-23
| |
* | merge in recent changes about param updatestxh182015-11-23
|\ \ | | | | | | | | | Merge branch 'master' into txh18/rnnlm
| * | small bug fixtxh182015-11-23
| |/
| * use consistent update calc; clean up code; no need for `direct_update`Determinant2015-11-21
| |
* | working on automatic parameter for layerstxh182015-11-20
|/
* added atomicAdd for select_linear update, however, the result still seems ↵txh182015-11-17
| | | | unreproducable, I changed select_linear layer update back to line-by-line
* change updateEI to update_by_err_inputtxh182015-11-16
|
* coding style changestxh182015-11-16
|
* unified param updates, now direct_update is the same speed with undirect_updatetxh182015-11-16
|
* used os.clock() for timertxh182015-11-16
|
* fixed direct update, did not know the resulttxh182015-11-16
|
* got good result when batch_size=1, strange!txh182015-11-11
|
* Merge pull request #6 from yimmon/masterTed Yin2015-10-12
|\ | | | | support batch_resize (sequence training related layers are in kaldi_seq)
| * kaldi mpe training supportYimmon Zhuang2015-09-18
| |
* | enable user to set direct_update in gconf; use direct_update in the exampleDeterminant2015-10-09
|/
* major change: use luarocks to manage projectDeterminant2015-06-22
Powered by cgit, Themed by Ted Yin.