All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
- Support revert permuting qk;
- Support permute qk to handle qk norm for Qwen3;
- Add extra check for MiniCPM to llama;
- Add subcommand
show-datato show tensor data;
- Replace logger from simple_logger 5.0 with flexi_logger 0.31;
0.3.0 - 2025-06-05
- Introduce glob to select multiple shard files using glob wildcards;
0.2.0 - 2025-02-24
- Add subcommand
diffto diff two gguf files; - Add q8 to f32 dequantize cast;
- Upgrade Rust to 2024 edition;
- Upgrade dependency
ggus0.4 to 0.5; - Format every file;