Skip to content

[4/8] fix: MLA RoPE triton kernel head indexing and v_dim=0 support - #5

Merged
yueming-yuan merged 1 commit into
miles-mainfrom
fix/mla-rope-triton-kernel
Feb 19, 2026
Merged

[4/8] fix: MLA RoPE triton kernel head indexing and v_dim=0 support#5
yueming-yuan merged 1 commit into
miles-mainfrom
fix/mla-rope-triton-kernel

Conversation

@yueming-yuan

Copy link
Copy Markdown
  • Fix absolute head index computation in fwd/bwd triton kernels
  • Add k_dim_ceil (power-of-2 alignment) for triton compatibility
  • Handle v_dim=0 case in both forward and backward kernels
  • Use proper mask_k/mask_v instead of single mask

Fix absolute head index computation in fwd/bwd triton kernels for
MLA YaRN RoPE. Add k_dim_ceil (power-of-2) for proper triton alignment
and handle v_dim=0 case gracefully.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
@yueming-yuan yueming-yuan changed the title fix: MLA RoPE triton kernel head indexing and v_dim=0 support [4/8] fix: MLA RoPE triton kernel head indexing and v_dim=0 support Feb 19, 2026
@yueming-yuan
yueming-yuan merged commit 5263201 into miles-main Feb 19, 2026
guapisolo pushed a commit that referenced this pull request Feb 26, 2026
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
guapisolo pushed a commit that referenced this pull request Mar 2, 2026
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant