The Unsloth team has updated all its Gemma-4 model uploads, incorporating an improved chat template and other potential enhancements. Users are recommended to re-download the models to benefit from these updates, which demonstrate Unsloth's ongoing support for the Gemma-4 launch.
New quantizations for the Unsloth MiniMax M2.7 model, ranging from Q1 to BF16, have been uploaded to Hugging Face. A detailed list of GGUF quantizations, along with their respective sizes, is now available for download.
Este conteúdo anuncia a atualização dos modelos Gemma 4 GGUF da Unsloth, incorporando várias melhorias e correções do projeto llama.cpp. As atualizações abordam aspectos técnicos como cache KV, suporte CUDA, manuseio de vocabulário e parsing específico para Gemma 4.
The files for the unsloth Qwen3.6-27B model in GGUF format are finally available. This update marks the release of the long-awaited files for the specified AI model.
