My solutions to the Unsloth Challenge, a set of open puzzles focused on optimizing LLM fine-tuning through Triton kernels, distributed training, compilation, and memory-efficient techniques.

Check the GitHub for more information.