view post Post 1117 We collaborated with NVIDIA to teach you how we made LLM training ~25% faster! πLearn how 3 optimizations help your home GPU train models faster:1. Packed-sequence metadata caching2. Double-buffered checkpoint reloads3. Faster MoE routingGuide: https://unsloth.ai/blog/nvidia-collabGitHub: https://github.com/unslothai/unsloth See translation π₯ 4 4 π€ 2 2 π 2 2 + Reply