Skip to content

DeepSpeed v0.3.1
Compare
Choose a tag to compare
@jeffra jeffra released this 12 Nov 19:55
· 2304 commits to master since this release
v0.3.1
31f46fe
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

Updates

  • Efficient and robust compressed training through progressive layer dropping
  • JIT compilation of C++/CUDA extensions
  • Python-only install support, ~10x faster install time
  • PyPI hosted installation via pip install deepspeed
  • Removed apex dependency
  • Bug fixes for ZeRO-offload and CPU-Adam
  • Transformer support for dynamic sequence length (#424)
  • Linear warmup+decay lr schedule (#414)
Assets 2
Loading