What's Changed
- sync upstream llama.cpp (b7179) by @ngxson in https://github.com/ngxson/wllama/pull/194
Full Changelog: https://github.com/ngxson/wllama/compare/2.3.6...2.3.7
What's Changed
- sync with llama.cpp upstream by @ngxson in https://github.com/ngxson/wllama/pull/192
Full Changelog: https://github.com/ngxson/wllama/compare/2.3.5...2.3.6
News
Firefox 142 now officially supports the wllama API for extensions 🚀
Link to release note: https://www.firefox.com/en-US/firef...
What's Changed
Small fix for KV cache management, which cause some issues with hybrid and recurrence models
- if KV rm fails, we should clear the whole cache by @ngxson in https://github.com/ngxson/wllama/pull/188
Full Changelog: https://github.com/ngxson/wllama/compare/2.3.3...2.3.4
What's Changed
With latest sync from llama.cpp, new models are now supported, including Hugging Face SmolLM3 and LiquidAI LFM2
- Strict tsconfig.json by @Jerboas86 in https://github.com/ngxson/wllama/pull/178
- Improve GGUF file validation to support URLs with query parameters by @felladrin in https://github.com/ngxson/wllama/pull/177
- fix ts strict mode by @ngxson in https://github.co...