claude
small-model-training
训练或微调小模型(<8B)的端到端交付工作流,训练引擎用 LLaMA-Factory。覆盖客户需求对齐、底座 bake-off 选型、数据工程与泄漏门禁、SFT 配方、冻结 golden 评测、错误分析驱动的迭代自省,以及一份 36+ 条的跨项目踩坑账本。
/
Plugin small-model-training from marketplace smt-skills in 860844175/skills-marketplace. Commands read from the repo's manifest — 1 month ago.
What it installs
Skills (1)
small-model-training训练或微调小模型(<8B)的端到端交付工作流,训练引擎 LLaMA-Factory。触发场景:想训模型(「我想训练模型」「训个小模型」「训一个模型做X」「帮我微调」「跑个 SFT/LoRA/QLoRA」「做蒸馏」「fine-tune a model」);选底座(用什么模型、多大够用、base/基座/底座选型);训练数据与评测集(要多少条、去哪找数据集、怎么标注、golden/测试集、数据泄漏、dataset_info);训练配置与排障(超参/lr/batch/epoch/LoRA rank/dropout/bf16/liger、OOM、loss 不降、看不到 loss、template、lla
Maintain small-model-training?
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.
claude & codex pluginslisted