Skill UI

Fine-tune LLMs with TRL’s RLHF toolkit, covering SFT, DPO, PPO/GRPO and reward model training so you can align HuggingFace Transformers with human preferences or feedback in post-training workflows.

A comprehensive skill for fine-tuning and training large language and vision models using advanced techniques like TRL (SFT, DPO, GRPO) or Unsloth. This process runs entirely on managed Hugging Face cloud GPUs, eliminating the need for local GPU setup. It supports modern training workflows, model conversion (GGUF for local deployment), and ensures seamless artifact storage to the Hugging Face Hub.

A comprehensive guide and workflow for deploying static websites (e.g., Vue, React, Angular, Hugo, Jekyll) from a local environment to a production server using VSCode SFTP. This skill covers setting up secure SFTP connections, generating production-ready Nginx configurations (including security headers, caching, and HTTPS redirects), and providing a complete deployment checklist for smooth CI/CD-like workflows.

This skill provides expert capabilities for training and fine-tuning transformer language models using the TRL (Transformers Reinforcement Learning) library. It supports state-of-the-art post-training methods, including Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), Group Relative Policy Optimization (GRPO), and Reward Model training. Use this skill to align and customize foundation models for advanced tasks via CLI commands.

Train or fine-tune language and vision models using TRL or Unsloth on Hugging Face Jobs infrastructure. Supports SFT, DPO, GRPO methods and GGUF conversion for local deployment. Requires Pro account and HF_TOKEN for Hub saving. Optimized for cloud GPU training without local setup.

Train and fine-tune transformer language models with Hugging Face TRL via CLI. Covers SFT, DPO, GRPO, RLOO, KTO and reward model training, plus LoRA adapters, YAML configs and distributed multi-GPU runs.

This skill prepares, formats, and validates datasets for supervised fine-tuning and preference training. It covers format selection for SFT, DPO, and RLVR, chat template application, sequence packing optimization, and synthetic data generation rules. It ensures data quality before training launch by creating dataset cards and validating loss masking.

Routes fine-tuning decisions: decide whether to fine-tune, choose SFT, DPO/ORPO/KTO, GRPO/RLVR, continued pretraining, and base model size. Helps avoid unnecessary training and select the right method.