awslabs/slapo

Issues

• Compatibility issues with DeepSpeed main branch

  #89 opened a year ago by zarzen
  3

• [Test] Unit test robustness

  #11 opened 2 years ago by comaniac
  2

• Enquiries about Parameter Sharding

  #98 opened a year ago by keneoneth
  3

• [Limitation] Flattened traced module cannot be consolidated

  #101 opened a year ago by chhzh123
  0

• [Feature] Type Inference of primitives and module selection API change

  #91 opened 2 years ago by zarzen
  7

• [Feature] End-to-end verification for module correctness

  #83 opened 2 years ago by chhzh123
  1

• [Bug] Multiple reductions of linear bias after row sharding the weight

  #10 opened 2 years ago by chhzh123
  0

• [Feature] A primitive to set `get_cuda_rng_tracker()`

  #67 opened 2 years ago by comaniac
  0

• [Tracer] PyTorch tracer does not rename SelfAttention module

  #30 opened 2 years ago by chhzh123
  2

• [Action] Failed to upload to PYPI

  #66 opened 2 years ago by comaniac
  0

• [Release] v0.0.2

  #47 opened 2 years ago by comaniac
  1

• [Bug] Buffer is not maintained and redundant buffer is returned after tracing bert example

  #13 opened 2 years ago by szhengac
  1

• [Bug] When `--checkpoint 0` is used, system raise RuntimeError: Tensors must be contiguous

  #43 opened 2 years ago by zarzen
  1

• [Feature] Random seed management for dropout layers in distributed environment

  #21 opened 2 years ago by chhzh123
  1

• [Feature] A systematic tensor parallelism verifier

  #37 opened 2 years ago by comaniac
  0

• [Bug] Original hooks are discarded in pipeline module

  #24 opened 2 years ago by chhzh123
  0

• [Bug] Sharing embedding weights with last linear layer

  #12 opened 2 years ago by zarzen
  1