Proto commits in apache/systemds

These 16 commits are when the Protocol Buffers files have changed:

2022-05-22

[SYSTEMDS-3381] Autogenerate java classes from proto definition with mvn * upgrade to protobuf 3.20.1 see https://github.com/protocolbuffers/protobuf/releases/tag/v3.20.1 * generate .java file * license exclude protobuf/*.java file * add a note on protoc binaries Closes #1620.

The documentation is generated from this commit.

2020-06-25

[SYSTEMDS-511] Add protobuf support Adds support for Protobuf file format, for both reads and write. AMLS project SS2020, part 1 Closes #971

2020-03-27

[SYSTEMDS-xxx] Merge SystemDS back into SystemML

This commit does not contain any .proto files.

2019-08-31

[MINOR] Fix missing license header (release prep rat check)

This commit does not contain any .proto files.

2019-03-23

[SYSTEMML-540] Added zero padding layer in Caffe2DML, Keras2DML and nn library - Updated the tests and the documentation. - This layer is required for ResNet-50 demo with Keras2DML.

2018-11-16

[MINOR] Removed various APIs, docs, and other resources

2018-09-07

Initial project commit, forked from Apache SystemML branch 1.2

2018-02-27

[SYSTEMML-1872] Added average pooling and upsampling layers - Added avg_pool and avg_pool_backward builtin functions. - The above builtin functions are supported on both CPU and GPU. - Also, added compare_backend tests for the above function to compare the results of CP operator with that of CuDNN. - Added avg_pool2d_builtin.dml and upsample2d.dml in the nn library. - Added gradcheck tests for the above layers. - Supported average pooling and upsampling in Keras2DML as well as Caffe2DML. - Tested the results of ResNet with average pooling on real-world images as a sanity check. - Also, tested upsampling layer by comparing the results with that returned by Keras. Closes #734.

2018-01-24

[SYSTEMML-540] Low-rank fully connected layer This layer has three advantages over the affine layer: - It has significantly lower memory requirement than affine layer making it ideal for devices such as GPUs. - It implicitly avoids overfitting by minimizing the number of parameters in the neural network. - It can exploit sparsity-aware fused operators. Closes #720.

2017-12-12

[SYSTEMML-540] Add support for RNN and LSTM in Caffe2DML and Keras2DML Closes #707.

2017-09-21

[SYSTEMML-1928] Removed Tensorboard visualization code from Caffe2DML - This avoids compilation error on powerpc. Closes #671.

2017-07-05

[SYSTEMML-540] Extended Caffe2DML to support image segmentation problems - This commit extends Caffe2DML to support image segmentation problem, depthwise convolution and has couple of bugfixes regarding loading existing caffe model. - Additionally, we have added a summary() method to Caffe2DML to print network. Closes #527.

2017-06-16

[SYSTEMML-540] Allow for `allreduce_parallel_batches` for multi-GPU training - Introduced `allreduce_parallel_batches` for multi-GPU training as per Mike's suggestion. - Moved `train_algo` and `test_algo` from solver specification to Python API to conform with Caffe as per Berthold's suggestion. - Updated the documentation for Caffe2DML. Closes #543.

2017-05-26

[SYSTEMML-1632] Support loading and saving models via mllearn - Also, updated documentation and fixed bugs.

2017-04-30

[SYSTEMML-1553] Add Caffe and Tensorflow Licenses based on discussion in the mailing list Closes #467.

2017-04-19

[SYSTEMML-692] Added initial version of DML generator for Caffe This experimental interface is called Caffe2DML and doesnot affect other functionality. - Updated the interface to match the Caffe specification as per @bertholdreinwald 's suggestion. - Added support for fine-tuning. - Added support for explain, statistics and gpu. Closes #422.