---

name: espnet2-training description: "Configure, dry-run, train, resume, fine-tune, and troubleshoot ESPnet2 task CLIs and PyTorch training workflows." disable-model-invocation: true metadata: disco-role: operating license: Apache 2.0

Espnet2 Training

Use this sub-skill for ESPnet2 task training modules, ConfigArgParse/YAML configuration, component choices, CPU dry-runs, checkpointing, fine-tuning, GPU/distributed flags, and training failures.

Quick use

• Read training-cli-reference.md, configuration-and-components.md, and gpu-distributed-training.md; use helper scripts for safe inspection.

Boundaries

Route data/recipe setup to ../recipes-and-data/SKILL.md, install/backend blockers to ../installation-and-diagnostics/SKILL.md, pretrained inference to ../inference-and-model-zoo/SKILL.md, and tests to ../development-and-testing/SKILL.md.

Safety

Do not run full recipes, model downloads, training jobs, uploads, demo servers, broad CI, or host-mutating installers without explicit user approval. Keep local/private environment details out of public answers.