Skip to main content
skill CX Codex

trainer-integration — Claude Code skill: features and install

trainer-integration is a Claude Code skill for adding or refining LoRA training support for models in Draw Things. It covers trainer wiring, checkpoint export, and validation.

GPL-3.0 ★ 171 updated 2026-11-17

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add drawthingsai/local-code-community --skill trainer-integration -a claude-code
Open on GitHub

What trainer-integration does

The skill lays out a workflow for adding LoRA training paths for generative models available in Draw Things app and CLI. It covers model builders, trainer dispatch, tokenizer setup, and fixed encoders. Flux1 serves as the reference path; the workflow advises reusing only the pieces a model needs.

The checklist covers trainable-key selection in both the CLI and app, weight loading with the required model key and codec list, rotary handling, and attention backend checks. It also addresses numerical stability and LoRA checkpoint export. Qwen Image uses the Qwen 2.5 tokenizer, while Z Image uses Qwen 3.

What it helps with

  • add a LoRA training path for a Draw Things model
  • align trainable-key selection between the CLI and app
  • connect the model's tokenizer and fixed encoder
  • check training numerics and LoRA checkpoint export

Example prompts for the agent

  • Add LoRA training for a new Draw Things model and use the Flux1 path as a reference.
  • Check which tokenizer the Qwen Image training path needs.
  • Find where to add version-specific model selection in the CLI and app.

How to install trainer-integration

1Skills CLI (Vercel)

npx skills add drawthingsai/local-code-community --skill trainer-integration -a claude-code

2Manually: put the files in place

~/.claude/skills/trainer-integration/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/trainer-integration/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about trainer-integration

How do I install trainer-integration in Claude Code?

Run npx skills add drawthingsai/local-code-community --skill trainer-integration -a claude-code. The skill file is installed at ~/.claude/skills/trainer-integration/SKILL.md.

What does trainer-integration do?

It guides work on LoRA training support for Draw Things models, including trainer wiring, tokenizer setup, numerical checks, and checkpoint export.

Which tokenizers do Qwen Image and Z Image use?

Qwen Image uses Qwen 2.5, while Z Image uses Qwen 3.

Similar skills

All skills for AI and ML →
scenario-model-training scenario-labs scenario-model-training — a skill for Claude Code that guides custom image and video model training. It activates for dataset curation, model selection, cost estimates, training runs, and generation through Scenario MCP. CX Codex +1 929installs llm-integration OrchestKit llm-integration — a collection of 20 production rules for LLM workflows: strict function calling, SSE streaming endpoints, local Ollama deployments, and LoRA tuning. CX Codex +1 218installs huggingface-llm-trainer huggingface huggingface-llm-trainer runs language and vision model fine-tuning on Hugging Face Jobs with TRL or Unsloth and converts checkpoints to GGUF without local GPUs. CX Codex +1 2,500installs huggingface-vision-trainer huggingface huggingface-vision-trainer — a skill that fine-tunes object detectors, image classifiers and SAM/SAM2 models on Hugging Face Jobs GPUs. Useful when you have no local GPU but still need a trained model on the Hub. CX Codex +1 2,100installs nemotron-customize NVIDIA Nemotron Team nemotron-customize is a Claude Code skill that plans Nemotron customization workflows. It activates for requests to train, evaluate, convert, or otherwise configure Nemotron models. CX Codex +1 2,200installs ml-model-training secondsky ml-model-training — a skill that walks the agent through the full model training cycle: data prep, splitting, training and evaluation. Useful when you need a classifier or regressor without skipping steps. CX Codex +1 484installs

From the same set20

The whole set →
deploy-static-site-with-wrangler deploy-static-site-with-wrangler is a Claude Code skill that uploads static site builds to temporary public Cloudflare URLs. It triggers when someone asks to share or preview built HTML assets. CX Codex +1 ★ 171repo stars train-lora train-lora is a Claude Code skill that validates LoRA training with Draw Things through draw-things-cli. It runs local dataset checks and compares generated images with and without the LoRA. CX Codex +1 ★ 171repo stars skill-creator skill-creator is a Claude Code skill that guides creating, validating, updating, and removing workspace skills. It triggers on requests for reusable workflows or tool integrations. CX Codex +1 ★ 171repo stars init-gpu-server init-gpu-server is a Claude Code skill that prepares a Draw Things GPU server by syncing GPUScript, installing server software, and mounting data disks. CX Codex +1 ★ 171repo stars run-longcat-avatar-video run-longcat-avatar-video is a Claude Code skill that orchestrates talking avatar generation via draw-things-cli. It animates a portrait from sound input in LongCat-Video-Avatar 1.5. CX Codex +1 ★ 171repo stars use-drawthings-cli use-drawthings-cli is a Claude Code skill for running and troubleshooting image generation with Draw Things CLI. It comes into play when choosing a model, setting up cloud access, or diagnosing a failed run. CX Codex +1 ★ 171repo stars setup-cpu-proxy-server setup-cpu-proxy-server is a Claude Code skill that deploys Draw Things CPU proxy and Envoy instances on Linux hosts. Runs during node provisioning via SSH or Tailscale. CX Codex +1 ★ 171repo stars new-model-integration new-model-integration is a Claude Code skill for adding image or video generation models to Draw Things. It guides the work from compilation through model wiring and CLI checks. CX Codex +1 ★ 171repo stars

Project sponsors

Companies that keep this analytics open