kernel.rst 653 B

1234567891011121314151617
  1. Transformer Kernels
  2. ===================
  3. The transformer kernel API in DeepSpeed can be used to create BERT transformer layer for
  4. more efficient pre-training and fine-tuning, it includes the transformer layer configurations and
  5. transformer layer module initialization.
  6. Here we present the transformer kernel API.
  7. Please see the `BERT pre-training tutorial <https://www.deepspeed.ai/tutorials/bert-pretraining/>`_ for usage details.
  8. DeepSpeed Transformer Config
  9. ----------------------------
  10. .. autoclass:: deepspeed.DeepSpeedTransformerConfig
  11. DeepSpeed Transformer Layer
  12. ----------------------------
  13. .. autoclass:: deepspeed.DeepSpeedTransformerLayer