Deploy the HMAI LoRA Trainer serverless endpoint on RunPod and train LoRAs end-to-end on it — SDXL, Wan 2.2, Qwen-Image (+2512), FLUX.2 Klein 9B, Krea 2 (+Turbo), Z-Image Turbo, and Ideogram 4. Use this skill whenever the user wants to train a LoRA, fine-tune a diffusion model on their own images, deploy the aio-lora-trainer template, set up the HMAI trainer endpoint, or asks anything about dataset prep, config overrides, checkpoints, or retrieving trained .safetensors from this endpoint — even if they just say "train a LoRA of X" without naming the service.
2026-07-08