Fine-Tune Transformer Model
curated by SkillMD · plugin · 8 skills
Fine-tune transformer language models using TRL with support for SFT, DPO, GRPO, and reward model training.
Install the whole plugin (CLI)
npx skillmds add huggingface/huggingface-best
npx skillmds add huggingface/trl-training
npx skillmds add affaan-m/eval-harness
npx skillmds add google-gemma/gemma-trainer
npx skillmds add majiayu000/dpo
npx skillmds add majiayu000/sft
npx skillmds add lord1egypt/peft-fine-tuning
npx skillmds add lord1egypt/simpo-trainingSkills in this plugin
- ▌ huggingface-best · huggingfaceQueries Hugging Face benchmark leaderboards to find the best AI models for a task, filters by device constraints, and returns a ranked comparison table with scores.
- ▌ trl-training · huggingfaceTrain and fine-tune transformer language models using TRL (Transformers Reinforcement Learning) with support for SFT, DPO, GRPO, KTO, RLOO, and reward model training via CLI commands.
- ▌ eval-harness · affaan-mProvides a formal evaluation framework for Claude Code sessions, implementing eval-driven development (EDD) principles to define pass/fail criteria, measure reliability with pass@k metrics, and create regression test suites.
- ▌ gemma-trainer · google-gemma bundleFine-tune Gemma models locally using QLoRA, Unsloth, or TRL for SFT, DPO, and reward modeling, with dataset preparation and conversion to GGUF or LiteRT-LM.
- ▌ dpo · majiayu000 bundleTrains language models with Direct Preference Optimization using preference pairs, covering DPOTrainer setup, dataset preparation, and beta tuning for stable preference learning without explicit reward models.
- ▌ sft · majiayu000 bundleFine-tune instruction-following LLMs with Unsloth's optimized SFTTrainer, covering dataset formatting, chat templates, training configuration, and thinking-model patterns.
- ▌ peft-fine-tuning · lord1egyptFine-tune large language models by training less than 1% of parameters using LoRA, QLoRA, and 25+ adapter methods, enabling efficient adaptation on limited GPU memory.
- ▌ simpo-training · lord1egyptTrains LLMs with SimPO, a reference-free preference optimization method that outperforms DPO, using configurable hyperparameters and workflows for various models and tasks.