协作头脑风暴
brainstorming
在实现前澄清需求、探索设计选择,并形成清晰的开发方向。
training-and-experiments【开发与工程】Run and debug H2O LLM Studio training experiments, distributed launchers, backend choices, and experiment artifacts.
独立 GitHub 快照 · 自动分类 · 以原作者文档为准
Use this sub-skill when the task is to launch, monitor, or debug an H2O LLM Studio experiment after a dataset/config has already been chosen. It owns direct CLI training, GUI-launched training behavior, multi-GPU launch mechanics, backend/performance settings, and the files written to an experiment output directory.
llm_studio/train.py -Y runs, deprecated -C, dynamic --section.field overrides, GUI launch behavior, and multi-GPU commands.scripts/make_minimal_config.py: create a tiny CSV plus CPU-like YAML config and print a non-executing training command.scripts/check_training_environment.py: inspect Python, package, PyTorch/CUDA, CUDA_HOME, nvcc, DeepSpeed importability, and a YAML config without starting training.scripts/distributed_train_wrapper.sh: build a validated torchrun or deepspeed multi-GPU command; dry-run by default and executes only with --execute.../configuration-and-data/SKILL.md.../modeling-and-evaluation/SKILL.md.../export-and-prompt/SKILL.md.../app-and-ui/SKILL.md.