DeepSpeed
Fixing several bugs in the inference-api and the kernels
#1951
Merged

Commits
  • fixing several bugs in the inference-api and the kernels
    Reza Yazdani committed 3 years ago
  • formatting
    jeffra committed 3 years ago
  • add hf inference unit test
    jeffra committed 3 years ago
  • update unit tests for fp32/16
    jeffra committed 3 years ago
  • skip test for now, needs debugging
    jeffra committed 3 years ago
  • Merge branch 'master' into fix-gpt-inference
    jeffra committed 3 years ago
  • Merge branch 'master' into fix-gpt-inference
    jeffra committed 3 years ago
  • fix rotary_pos embedding for GPT-J inderence
    Reza Yazdani committed 3 years ago
  • Merge branch 'fix-gpt-inference' of github.com:microsoft/DeepSpeed into fix-gpt-inference
    Reza Yazdani committed 3 years ago
  • turn unit test back on
    jeffra committed 3 years ago
  • fix for gpt-j fp16 ckpt revision
    jeffra committed 3 years ago
  • add injection workaround for pt1.2
    jeffra committed 3 years ago
  • fix for amd
    jeffra committed 3 years ago
  • fix for megatron + pt1.2
    jeffra committed 3 years ago
  • skip on torch 1.2, several issues there
    jeffra committed 3 years ago
Loading