v0.4.1
davidfowl/tallyv0.4.1Nov 10, 2025by github-actions[bot]
AI Summary
This patch release fixes a critical memory issue on Blackwell GPUs where the RoPE (Rotary Positional Embedding) cache was expanding excessively, causing crashes.
Key Highlights
- Fixes memory blow-up on Blackwell GPUs
- Resolves bogus RoPE expansion issue (cache size ~4e18 positions)
Full Release Notes
## Bug Fixes - avoid bogus RoPE expansions on Blackwell GPUs https://github.com/TimmyOVO/deepseek-ocr.rs/issues/24 Blackwell builds were blowing up memory because the RoPE cache tried to grow to ~4e18 positions.