chutes

unsloth/Mistral-Nemo-Instruct-2407

Mistral Nemo Instruct 2407 bf16

Unsloth is a finetuning framework that enables 2-5x faster training of open-source LLMs (Llama, Mistral, Gemma) with 70% less memory usage, useful for efficiently adapting pre-trained models to custom datasets.