Age | Commit message (Collapse) | Author | |
---|---|---|---|
2016-09-27 | making ClassNLLCriterion targets consistent between cpu and cudaclassnllfix | soumith | |
2016-09-24 | Merge pull request #976 from Atcold/master | Soumith Chintala | |
Remove extra | and bug fix ConcatTable | |||
2016-09-22 | Remove extra | and bug fix ConcatTable | Alfredo Canziani | |
2016-09-20 | Merge pull request #973 from jgehring/clone-share | Soumith Chintala | |
Don't serialize shared parameters in clone() | |||
2016-09-20 | Don't serialize shared parameters in clone() | Jonas Gehring | |
There's no need to serialize parameters that will be shared immediately afterwards. In particular, the current implementation of clone and share is inefficient for single modules in a larger network with flattened parameters since the storage holding all parameters will be serialized. Here, the write method of nn.Module is temporarily overwritten with a function that writes empty tensors for all shared parameters. | |||
2016-09-20 | Merge pull request #971 from hughperkins/migrate-optim-to-optim | Soumith Chintala | |
Migrate optim to optim | |||
2016-09-20 | migrate optim example to optim repo | Hugh Perkins | |
2016-09-20 | add StochasticGradient doc back | Hugh Perkins | |
2016-09-14 | Merge pull request #967 from apaszke/THNN_format_fix | Soumith Chintala | |
Mark BCECriterion weights as optional in THNN.h | |||
2016-09-14 | Mark BCECriterion weights as optional in THNN.h | Adam Paszke | |
2016-09-14 | adding storageOffset in VolumetricConvolutionMM | Soumith Chintala | |
2016-09-14 | Merge pull request #965 from apaszke/mm_weight | Soumith Chintala | |
Accept 5D weights in VolumetricConvolutionMM | |||
2016-09-14 | Accept 5D weights in VolumetricConvolutionMM | Adam Paszke | |
2016-09-13 | Merge pull request #964 from wydwww/master | Soumith Chintala | |
Just typos | |||
2016-09-13 | Just typos | Yiding Wang | |
2016-09-08 | Merge pull request #626 from szagoruyko/bcethnn | Soumith Chintala | |
BCECriterion THNN implementation | |||
2016-09-02 | fix bilinear doc | Soumith Chintala | |
2016-09-02 | Stochastic Inference for Dropout (#936) | Jason Kuen | |
* Stochastic Inference for Dropout | |||
2016-09-02 | Merge pull request #948 from culurciello/patch-1 | Soumith Chintala | |
Typo in doc links | |||
2016-09-02 | Typo in doc links | Eugenio Culurciello | |
2016-08-31 | Merge pull request #946 from gchanan/HardTanhBoundary | Soumith Chintala | |
HardTanh does not use inclusive bounds in inplace mode | |||
2016-08-31 | HardTanh does not use inclusive bounds in inplace mode | Gregory Chanan | |
HardTanh does not use inclusive bounds in inplace mode, which makes it inconsistent with the cuda version, as well as Threshold, ReLU, etc. | |||
2016-08-27 | Merge pull request #943 from kmul00/voldilmaxpool | Soumith Chintala | |
Add Volumetric Dilated Max Pooling | |||
2016-08-28 | Add Volumetric Dilated Max Pooling | kmul00 | |
new file: VolumetricDilatedMaxPooling.lua modified: init.lua modified: lib/THNN/generic/THNN.h copied: lib/THNN/generic/VolumetricMaxPooling.c -> lib/THNN/generic/VolumetricDilatedMaxPooling.c modified: lib/THNN/generic/VolumetricMaxPooling.c modified: lib/THNN/init.c modified: test.lua Update docs modified: doc/convolution.md Resolved bug in ceil mode modified: lib/THNN/generic/VolumetricDilatedMaxPooling.c | |||
2016-08-27 | Merge pull request #941 from kmul00/consistentmaxpoolapi | Soumith Chintala | |
Make SpatialMaxPooling API consistent | |||
2016-08-25 | Make SpatialMaxPooling API consistent | kmul00 | |
modified: SpatialDilatedMaxPooling.lua modified: SpatialMaxPooling.lua renamed: lib/THNN/generic/SpatialMaxPooling.c -> lib/THNN/generic/SpatialDilatedMaxPooling.c modified: lib/THNN/generic/SpatialMaxPooling.c modified: lib/THNN/generic/THNN.h modified: lib/THNN/init.c | |||
2016-08-25 | add CMaxTable class (#885) | Jonathan Uesato | |
* Add CMaxTable and CMinTable | |||
2016-08-25 | Adding TemporalDynamicKMaxPooling (#867) | Multihuntr | |
* Adding TemporalDynamicKMaxPooling | |||
2016-08-25 | Merge pull request #940 from torch/jtfix | Soumith Chintala | |
fixing JoinTable bug for negative indices | |||
2016-08-25 | fixing JoinTable bug for negative indices | soumith | |
2016-08-25 | Expand CMul weights (#911) | nhynes | |
* Expand CMul weights | |||
2016-08-22 | Merge pull request #937 from jgehring/maptable | Soumith Chintala | |
A container for applying a single module to all table elements | |||
2016-08-22 | fixing critical bug in SpatialConvolution | soumith | |
2016-08-22 | A container for applying a single module to all table elements | Jonas Gehring | |
This module can be useful when dealing with multi-dimensional mini-batches, for example when training recurrent neural networks (time X batch_size X elements). It simply clones the contained module as necessary and applies each clone to the respective input element. | |||
2016-08-20 | Merge pull request #935 from apaszke/threshold | Soumith Chintala | |
Make Threshold THNN functions more consistent | |||
2016-08-19 | Make Threshold THNN functions more consistent | Adam Paszke | |
2016-08-18 | Merge pull request #934 from apaszke/spatial_conv | Soumith Chintala | |
Accept both 2D and 4D weights in SpatialConvolutionMM | |||
2016-08-18 | Accept both 2D and 4D weights in SpatialConvolutionMM | Adam Paszke | |
2016-08-17 | Merge pull request #933 from torch/maxminfix | Soumith Chintala | |
fixing Max and Min for new cutorch types | |||
2016-08-17 | Merge pull request #932 from jgehring/concat-size | Soumith Chintala | |
Don't overwrite self.size in Concat | |||
2016-08-17 | fixing nn.Normalize for new cutorch types | soumith | |
2016-08-17 | fixing Max and Min for new cutorch types | soumith | |
2016-08-17 | Don't overwrite self.size in Concat | Jonas Gehring | |
size() is a method of the Container base class. | |||
2016-08-13 | Merge pull request #929 from CDLuminate/fix-spelling-errors | Soumith Chintala | |
Fix many spelling errors with tool `codespell`. | |||
2016-08-13 | Fix many spelling errors with tool `codespell`. | Zhou Mo | |
This patch is imported from Debian Package. | |||
2016-08-13 | Merge pull request #928 from szagoruyko/fix-test-typos | Soumith Chintala | |
Fix typos in tests | |||
2016-08-13 | fix typos in tests | Sergey Zagoruyko | |
2016-08-12 | Merge pull request #925 from kmul00/master | Soumith Chintala | |
SpatialDilatedMaxPooling doc | |||
2016-08-12 | SpatialDilatedMaxPooling doc | kmul00 | |
modified: doc/convolution.md | |||
2016-08-12 | Add Spatial Dilated Max Pooling | kmul00 | |
new file: SpatialDilatedMaxPooling.lua modified: SpatialMaxPooling.lua modified: init.lua modified: lib/THNN/generic/SpatialMaxPooling.c modified: test.lua Updated THNN modified: THNN.h Removed unnecessary function calls modified: SpatialDilatedMaxPooling.lua |