Skip to main content
Run any Skill in Manus
with one click

metis-finetune-eval

Stars0
Forks0
UpdatedJuly 14, 2026 at 08:08

End-to-end playbook to fine-tune and evaluate an LLM on ThakiCloud Metis (kubeflow-llm-training + vLLM serving on compute-h200). Covers served-model data generation, datasetโ†’S3, TrainJob submission (with every admission gotcha), LoRA adapter serving for inference, and gate-based A/B eval. Use when "๋ชจ๋ธ ํ•™์Šต์‹œ์ผœ์ค˜", "ํŒŒ์ธํŠœ๋‹ ๋Œ๋ ค์ค˜", "TrainJob ๋„์›Œ์ค˜", "ํ•™์Šต ์ „ํ›„ ํ‰๊ฐ€/A/B", "๋ฐ๋ชจ GPU๋กœ ํ•™์Šต", "fine-tune on the cluster", "train and evaluate", "serve the adapter and compare". General across models. Do NOT use for local/Colab training (use colab-unsloth-finetune) or serving-only ops (use demo-llm-inference-test).

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

File Explorer
3 files
SKILL.md
readonly