aboutsummaryrefslogtreecommitdiff
path: root/nerv/layer
Commit message (Collapse)AuthorAge
* complete bp of softmax layerQi Liu2016-06-21
|
* add bp function for softmax layerQi Liu2016-06-21
|
* change RNN/LSTM/LSTMP parameter order, which is compatible to old versionQi Liu2016-05-24
|
* lstm & lstmp layer support multiple inputQi Liu2016-05-12
|
* clip on gradient nowQi Liu2016-05-09
|
* fix bug in affine.luaQi Liu2016-05-09
|
* merge affine & lstm_gate & projection layerQi Liu2016-05-09
|
* prepare for the replacement of `asr_trainer.lua` with `trainer.lua`Determinant2016-05-08
|
* Merge branch 'master' of ssh://speechlab.sjtu.edu.cn:8022/nerv-dev/nervbeta-1Determinant2016-05-05
|\
| * fix param_repo:add bugQi Liu2016-05-05
| |
* | make `bind_params` and `init` consistent in `affine.lua`,Determinant2016-05-05
|/ | | | `lstm_gate.lua`, `projection.lua`
* add `no_update` attribute to `nerv.MatrixParam`; add correspondingDeterminant2016-05-03
| | | | settings to `nerv.AffineLayer` and `nerv.LSTMGateLayer`
* add relu layerQi Liu2016-04-29
|
* back-propagate error to the second head of nerv.MSELayerDeterminant2016-04-28
|
* ...Determinant2016-04-28
|
* add docDeterminant2016-04-28
|
* add projection layer and lstmp layertan tian2016-04-18
|
* support SeqBuffer decoding; fix bugs in finding paramsalpha-4Determinant2016-04-18
|
* fix bug in passing err_input to network; gen zero vectors for biasalpha-3.3Determinant2016-03-30
|
* fix bug of momentum & update mse layerQi Liu2016-03-29
|
* fix "not implemented" and lstm rebinding bugsalpha-3.1Determinant2016-03-26
|
* RNN support multiple input & fix network flush bugQi Liu2016-03-17
|
* add mask on softmax_ceQi Liu2016-03-15
|
* speedup duplicate layerQi Liu2016-03-15
|
* fix bug of dropoutQi Liu2016-03-11
|
* update diagonlal lstmQi Liu2016-03-11
|\
| * fix bugQi Liu2016-02-24
| |
| * fix typoQi Liu2016-02-24
| |
| * update LSTM layerQi Liu2016-02-24
| |
* | add recipe and fix bugsQi Liu2016-03-11
| |
* | adapt asr_trainer.lua to new architectureDeterminant2016-03-11
| |
* | Merge branch 'master' of https://github.com/liuq901/nerv into liuq901-masterDeterminant2016-03-11
|\ \ | | | | | | | | | | | | | | | Conflicts: nerv/layer/init.lua nerv/nn/layer_repo.lua
| * | update lstmQi Liu2016-03-04
| | |
| * | update add_prefix for graph layerQi Liu2016-03-03
| | |
| * | fix bugQi Liu2016-03-02
| | |
| * | add identity layerQi Liu2016-03-02
| | |
| * | network init completeQi Liu2016-03-01
| | |
| * | add network & fix graph_layerQi Liu2016-02-29
| | |
| * | graph layer completeQi Liu2016-02-29
| |/
* / major change: clearer param binding semantics; permit rebinding; enableDeterminant2016-03-10
|/ | | | resuming from previous training
* clean up obsolete filesDeterminant2016-02-22
|
* add layers from `layersT/` to `layer/`Determinant2016-02-21
|
* generalize softmax_ce.lua (to softmax_ce_t.lua)Determinant2016-02-18
|
* fixed some minor problemtxh182016-02-16
|
* ...Determinant2016-02-16
|
* improve parameter auto-detectionDeterminant2016-02-16
|
* move curand_gen out of the matrix layoutDeterminant2016-02-15
|
* try a basic mergeDeterminant2016-02-15
|
* changed affine_recurrent used in lm to affine_recurrent_plusvectxh182015-12-23
|
* trying to use lstm_t_v2 for ptbtxh182015-12-04
|
Powered by cgit, Themed by Ted Yin.