hpo-experiments
hpo-experiments
【效率与协作】Use NNI HPO experiments, nnictl, ExperimentConfig, trial metric APIs, search spaces, tuners, assessors, and training-service config triage.
THE SKILLS DIRECTORY
探索全部技能.
一个技能，一个新的可能。找到适合你的 AI 工作流。
peft
peft
【开发与工程】Use PEFT for parameter-efficient fine-tuning adapters, LoRA and quantized workflows, prompt and soft methods, specialized tuners, save/load/merge operations, training integrations, and PEFT repository development.
search-and-export
search-and-export
【效率与协作】Configure AutoKeras search, tuners, persistence, callbacks, export_model, and saved Keras model reloads.
specialized-tuners
specialized-tuners
【效率与协作】Select, configure, and troubleshoot PEFT's specialized non-LoRA/non-prompt tuners such as IA3, BOFT, OFT, LyCORIS-style LoHa/LoKr, VeRA/PVeRA, VBLoRA, XLora, LN tuning, FourierFT, FRoD, HRA/HiRA/SHiRA, RandLoRA, RoAd, OSF, and newer experimental tuner configs.