Skip to content

v0.5.6

Choose a tag to compare

View all tags
@github-actions github-actions released this 16 Jul 13:22
v0.5.6
4e46d7a
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode. 
v0.5.6: half precision (bf16/fp16) + opt-in lazy sync

- bf16 + fp16 first-class across all ops; cublasGemmEx tensor-core GEMM
  with fp32 accumulate; autocast integration (PR #7)
- ME_LAZY_SYNC=1 opt-in: skips 11 redundant per-op stream syncs,
  ~7-11% training step-time reduction at small batch (PR #8)
- Trench maintenance: NVTX/dead-header prune, Python dispatch cache (PR #6)
- Wheels: cu128/sm_120 CUDA builds, torch 2.9 line (PRs #5, #9, #10)
- CUDA 12.8 / Blackwell (sm_120), torch >= 2.7 (wheels: 2.9.x), NumPy >= 2.0
Assets 5
Loading