ml-ane-transformers

Reference implementation of the Transformer architecture optimized for Apple Neural Engine (ANE)

OTHER License

Downloads
1.3K
Stars
2.5K

Bot releases are visible (Hide)

ml-ane-transformers - v0.1.3 Latest Release

Published by atiorh about 2 years ago

ml-ane-transformers - v0.1.2

Published by atiorh about 2 years ago

  • torch.nn.LayerNorm and ane_transformers.reference.layer_norm.LayerNormANE apply scale and bias terms in opposite orders. In order to accurately restore a state_dict trained using the former into the the latter, we adjust the bias term. This change slightly improves the parity between the Hugging Face PyTorch model's outputs and ane_transformers CoreML model's outputs.
  • Introduce upper bound for torch version at 1.11.0 to ensure that coremltools-5.2.0 is well paired. The upper bound will be removed/updated when coremltools-6.0 is released (currently in beta1).
ml-ane-transformers - v0.1.1

Published by atiorh over 2 years ago