b7620
📦 llama-cppView on GitHub →
🐛 1 fixes🔧 2 symbols
Summary
This release (b7620) focuses on a fix for token padding reservation relative to the number of sequences in the context.
🐛 Bug Fixes
- Fixed reserve token padding to n_seqs in context management (#18536)
🔧 Affected Symbols
llama_contextn_seqs