MIOpen
MIOpen copied to clipboard
AMD's Machine Intelligence Library
* Added CosineEmbeddingLoss forward and backward operation and kernel. * Added driver test and gtest for CosineEmbeddingLoss. * New API is guarded by MIOPEN_BETA_API macro. * Average over all cases:...
* Added [MarginRankingLoss](https://pytorch.org/docs/stable/generated/torch.nn.MarginRankingLoss.html) forward and backward. * Added driver test and gtest for MarginRankingLoss. * New API is guarded by MIOPEN_BETA_API macro. * Average over all cases: | Type |...
* Added [SGD ](https://pytorch.org/docs/stable/generated/torch.optim.SGD.html) forward and backward. * Added driver test and gtest for SGD. * New API is guarded by MIOPEN_BETA_API macro. * Average over all cases: * SGD...
* Added no reduction [KLDivLoss ](https://pytorch.org/docs/stable/generated/torch.nn.KLDivLoss.html)backward operation and kernel. * Added driver test and gtest for KLDivLoss. * New API is guarded by MIOPEN_BETA_API macro. * Average over all cases:...
* Added [FractionalMaxpool2d](https://pytorch.org/docs/stable/generated/torch.nn.FractionalMaxPool2d.html) and [FractionalMaxpool3d](https://pytorch.org/docs/stable/generated/torch.nn.functional.fractional_max_pool3d.html). * Added driver test and gtest for FractionalMaxpool. * New API is guarded by MIOPEN_BETA_API macro. * Average over all cases: * FractionalMaxpool 2d |...
* Added LPPool [1D](https://pytorch.org/docs/stable/generated/torch.nn.LPPool1d.html) [2D](https://pytorch.org/docs/stable/generated/torch.nn.LPPool2d.html) forward and backward. * Added driver test and gtest for LPPool. * New API is guarded by MIOPEN_BETA_API macro. * Average over all cases: LPPool1D...
* Added AdaptiveMaxPool [1D](https://pytorch.org/docs/stable/generated/torch.nn.AdaptiveMaxPool1d.html) [2D](https://pytorch.org/docs/stable/generated/torch.nn.AdaptiveMaxPool2d.html) [3D](https://pytorch.org/docs/stable/generated/torch.nn.AdaptiveMaxPool3d.html) forward and backward. * Added driver test and gtest for AdaptiveMaxPool . * New API is guarded by MIOPEN_BETA_API macro. * Average over all...
* Added AvgPool [2D](https://pytorch.org/docs/stable/generated/torch.nn.AvgPool2d.html) [3D](https://pytorch.org/docs/stable/generated/torch.nn.AvgPool3d.html) forward and backward with new solvers and bfp16 supported. * Added driver test and gtest for AvgPool. * New API is guarded by MIOPEN_BETA_API macro....
This PR is a continuation of PR #3182. Accidently, I have closed the older PR, and then made change to the working branch, which makes me cannot reopen the older...
* Added SmoothL1Loss forward and backward. * Added driver test and gtest for both direction of SmoothL1Loss. * New API is guarded by MIOPEN_BETA_API macro. * Compared to ROCm pytorch:...