2 Commits

Author	SHA1	Message	Date
Luke Yeager	a47652379f	Fix SparseAdagrad for indices.ndim>1 ... Summary: Same fix as https://github.com/caffe2/caffe2/pull/249, but for SparseAdagrad. Also update the tests for both ops to test this functionality. Closes https://github.com/caffe2/caffe2/pull/675 Differential Revision: D5148750 Pulled By: akyrola fbshipit-source-id: d30b722429bc547fd53400c1a29e4ee9e2e6ed18	2017-05-30 12:02:18 -07:00
Luke Yeager	8bd0522c20	Add tests and GPU impls for sparse optimizers ... Summary: These GPU paths are probably even buggier than the CPU paths for sparse gradients with duplicate indices. Both paths cause multiple momentum updates in a single iteration, but only the GPU path is non-deterministic. Depending on how we decide to address the issues on the CPU path, pooyadavoodi has a good idea for how to match dense behavior with the sparse GPU ops. Closes https://github.com/caffe2/caffe2/pull/254 Reviewed By: bwasti Differential Revision: D4871680 Pulled By: dzhulgakov fbshipit-source-id: 220be57a0f699a22ea85ed4f7022d92d362d06b3	2017-04-13 11:07:40 -07:00