github facebookresearch/xformers v0.0.12

latest releases: v0.0.29.post3, v0.0.29.post2, v0.0.29.post1...
2 years ago

[0.0.12] - 2022-08-08

Fixed

  • Removed duplicated biases in the FusedMLP layers [#317]
  • Rotary embeddings respecting input types [#326]
  • Poolformer style instantiating useless projection layers [#349]
  • Fix layer position not being properly tracked, causing extra layernorms for programmatic xformers [#348]
  • Pass use_triton flag to LayerNorm module [#336]

Added

  • Four blocksparsity layouts from DeepSpeed [#320]
  • Support several initialization options [#312]
  • Conv2DFeedforward feedforward part [#321]
  • VisualAttention [#329]
  • Automatic blocksparse for causal attention [#334]
  • Better hierarchical transformer generation [#345]
  • Fused operations with AOTAutograd/NVFuser, integration into MLP [#357]
  • Refactor LRA code to use Pytorch Lightning [#343]

Don't miss a new xformers release

NewReleases is sending notifications on new releases.