training-and-evaluation
Compose and debug Diffusion Policy Hydra training, evaluation, Ray multirun, checkpoint evaluation, and benchmark logging commands.
- License 许可
- MIT
- 最近更新
- 2026-10-08
- Stars
- ★ 11
- Forks
- 3
01 前言 Front Matter
- name
- training-and-evaluation
- license
- MIT
- metadata
- { "disco-role": "operating" }
- description
- Compose and debug Diffusion Policy Hydra training, evaluation, Ray multirun, checkpoint evaluation, and benchmark logging commands.
- disable-model-invocation
- true
02 资源文件 3
- 参考文档references/configuration.md5.7 KB
- 参考文档references/training-and-evaluation.md7.9 KB
- 参考文档references/troubleshooting.md6.7 KB
03 SKILL.md 正文
Training and Evaluation
Use this sub-skill when you need to build or debug Diffusion Policy training and evaluation workflows.
Route here for
- Single-seed Hydra training from
train.py. - Checkpoint evaluation from
eval.py. - Ray multirun orchestration from
ray_train_multirun.pyandray_exec.py. - Offline aggregation of
train_*/logs.json.txtfrommultirun_metrics.py-style logs. - Simulation benchmark routing for low-dim, image, hybrid, Robomimic, and video workspaces.
- W&B, checkpoint, and output-tree diagnostics for training/evaluation runs.
Do not handle here
- Dataset zarr schema, conversion, replay-buffer layout, or sampling details: route to data-and-replay-buffers.
- Policy/model architecture, losses, normalizers, or inference internals: route to policies-and-models.
- UR5, RealSense, SpaceMouse, or live robot execution: route to real-robot-operations.
Operating workflow
- Choose the workspace config family and matching task config.
- Compose the resolved config with
scripts/compose_experiment_config.py. - Launch a single-seed
train.pyrun or a Ray multirun. - Inspect
logs.json.txt, checkpoints,eval_log.json, andmetrics/logs.json.txt. - Use
scripts/summarize_multirun_metrics.pyto summarize finishedtrain_*logs offline.
Core API names
BaseWorkspace.run()drives the training/evaluation lifecycle.BaseWorkspace.save_checkpoint()andBaseWorkspace.load_checkpoint()manage checkpoint state.policy.set_normalizer(...)andpolicy.predict_action(...)are the key policy hooks.env_runner.run(policy)produces rollout metrics and videos.JsonLogger,read_json_log, andTopKCheckpointManagershape the run logs.
Start here
Read the bundled references for command recipes, config structure, and troubleshooting:
04 同仓库更多 Skills
obsidian-clipper-template-creator
Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.
obsidian-clipper-template-creator
Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.
python-api
Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP/mAP semantics, and source-style import diagnostics.
file-evaluation
Folder-based PASCAL VOC AP/mAP evaluation from ground-truth and detection text files.
object-detection-metrics
Guides legacy Object-Detection-Metrics PASCAL VOC AP/mAP evaluation from detection text folders and direct Python API classes.
nx-workspace-patterns
Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.