Commits · aed2f75e209e525c842aec7c044af7acae2a4614 · wxj / Megatron-LM

"sgl-kernel/csrc/allreduce/test_mscclpp_allreduce.cu" did not exist on "4474eaf5528aa073ce5ea6dc8c4136dc2b8f7449"

19 Mar, 2021 1 commit
- Bfloat fused softmax + fused layer norm · 0fa7175f
  Mohammad Shoeybi authored Mar 19, 2021
  
  0fa7175f
18 Mar, 2021 1 commit
- refactored the fused kernels build · 0d5188c1
  Mohammad Shoeybi authored Mar 17, 2021
  
  0d5188c1
09 Jan, 2021 1 commit
- decoder support in transformers · 4b506832
  Vijay Korthikanti authored Jan 08, 2021
  
  4b506832
18 Sep, 2020 1 commit
- scaling-generalized masking-softmax fusion kernel · d73def35
  root authored Sep 18, 2020
  
  d73def35
11 Sep, 2020 1 commit
- Various speed optimizations. · 51a2e6b0
  Vijay Korthikanti authored Sep 11, 2020
  
  51a2e6b0