* Implementations of `scatter_mul` and `segment_mul` * Reduced the size of shared library files * CUDA wheels can now also operate on CPU-only devices
2.0.5
* Fixed a bug in which the backward pass of `scatter_std` did not succeed
2.0.4
Fixed a bug where `composite` functions could not be differentiated.
2.0.3
* Windows fixes * Wheels for common OS/CUDA combinations
2.0.2
**New major release** including * fully-traceable operators * (deterministic) "scatter" operations based on sorted indices: `segment_csr` and `segment_coo`
**PyTorch 1.4 is now required** to install this package.
1.4.0
* Added `scatter_logsumexp` implementation * Added new **.composite** subpackage with `scatter_softmax` and `scatter_log_softmax` implementations