Skip to content
cccskills
FreePublished on GitHub

di-engine

Routes DI-engine users to the right workflow for CLI/config launches, serial and framework-based RL pipelines, and environment integration.

The English summary is being prepared. Showing the original description.

See how to install

Files included (26)

  • SKILL.md3.3 KB
  • references/package-surface.md2.3 KB
  • references/repo-provenance.md1.0 KB
  • references/repo-routing-metadata.json520 B
  • references/troubleshooting.md3.1 KB
  • scripts/check_install.py1.2 KB
  • scripts/run_parallel.sh197 B
  • scripts/run_serial.sh195 B
  • sub-skills/cli-config/references/cli-reference.md2.7 KB
  • sub-skills/cli-config/references/config-reference.md2.2 KB
  • sub-skills/cli-config/references/troubleshooting.md1.3 KB
  • sub-skills/cli-config/scripts/cli_smoke.py1.1 KB
  • sub-skills/cli-config/SKILL.md2.9 KB
  • sub-skills/env-integration/references/environment-reference.md2.6 KB
  • sub-skills/env-integration/references/troubleshooting.md1.7 KB
  • sub-skills/env-integration/scripts/env_wrapper_smoke.py1.1 KB
  • sub-skills/env-integration/SKILL.md2.8 KB
  • sub-skills/framework-runtime/references/api-reference.md2.6 KB
  • sub-skills/framework-runtime/references/troubleshooting.md1.4 KB
  • sub-skills/framework-runtime/references/workflows.md2.1 KB
  • sub-skills/framework-runtime/scripts/parallel_smoke.py781 B
  • sub-skills/framework-runtime/scripts/task_smoke.py694 B
  • sub-skills/framework-runtime/SKILL.md2.7 KB
  • sub-skills/serial-pipelines/references/troubleshooting.md1.6 KB
  • sub-skills/serial-pipelines/references/workflows.md2.6 KB
  • sub-skills/serial-pipelines/SKILL.md2.9 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

DI-engine

Use this skill for the public DI-engine reinforcement-learning engine. It routes common user requests to focused sub-skills instead of forcing future agents to rediscover the repo structure from scratch.

Quick check

If you only need to verify that the package is installed, start with:

python -m pip check
python -c "import ding, dizoo; print(ding.__version__)"

For a faster bundled smoke check, run scripts/check_install.py from an environment where DI-engine is installed.

Main routes

cli-config

Read this when the user asks about:

  • ding or ditask CLI usage
  • config files, main_config, create_config, or system_config
  • compile_config, read_config, save_config, or registry queries
  • launch flags for serial, parallel, dist, eval, or ditask
  • the repo-provided shell wrappers for serial/parallel launches

serial-pipelines

Read this when the user asks about:

  • ding.entry.serial_pipeline, serial_pipeline_onpolicy, or serial_pipeline_offline
  • eval, collect_demo_data, collect_episodic_demo_data, or episodic post-processing
  • legacy ding/entry/* launchers and environment-specific entry scripts under dizoo/
  • special modes such as SQIL, GAIL, DQFD, TREX, NGU, reward-model, or offline RL

framework-runtime

Read this when the user asks about:

  • ding.framework.task, Parallel, Supervisor, or EventLoop
  • middleware-based recipes using StepCollector, OffPolicyLearner, multistep_trainer, interaction_evaluator, CkptSaver, or data_pusher
  • the modern example scripts in ding/example/
  • multi-process message routing, task.wait_for, task.emit, or auto-recovery

env-integration

Read this when the user asks about:

  • DingEnvWrapper, BaseEnvManager, BaseEnvManagerV2, or EnvSupervisor
  • env wrapper selection, reset/step shape mismatches, or new environment onboarding
  • DIZoo environment patterns in dizoo/classic_control/, bitflip/, frozen_lake/, or league_demo/
  • custom environment adapters and the env-specific config/template pattern

Useful bundled helpers

  • scripts/check_install.py — quick import/version/config smoke for a prepared environment.
  • scripts/run_serial.sh — thin wrapper for ding -m serial with a config path and seed.
  • scripts/run_parallel.sh — thin wrapper for ding -m parallel with a config path and seed.

Read first

  • references/package-surface.md for the repo module map and public entry points.
  • references/troubleshooting.md for cross-cutting install/import/runtime failures.

Notes

  • This skill covers the core DI-engine workflows that can be exercised from the installed package and bundled helpers without reopening the source checkout.
  • Advanced external environment families that need separate optional packages are intentionally left out of the default routes.
  • If a request mentions a specific example, config, or entry script, route to the sub-skill that owns that workflow rather than reading the source tree directly.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Routes 3D ResNets PyTorch video action-recognition workflows across training, inference, and data preparation.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill339Updated Sep 3, 2026

3ddfa

Free

Guide 3DDFA Python inference, geometry rendering, training/evaluation, and optional C++ ONNX workflows for 3D dense face alignment.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill339Updated Sep 3, 2026

Routes 3DDFA_V2 face-alignment setup, still-image demos, video tracking, and ONNX benchmarking workflows.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill339Updated Sep 3, 2026

Operate AB3DMOT 3D multi-object tracking workflows for KITTI and nuScenes data, tracking, evaluation, and visualization.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill339Updated Sep 3, 2026

Use Hugging Face Accelerate for PyTorch training-loop migration, distributed launch/configuration, DeepSpeed/FSDP/TPU backend setup, big-model inference/offload, checkpointing, tracking, and troubleshooting.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill339Updated Sep 3, 2026

acme

Free

Route Acme reinforcement-learning framework tasks across core loops, replay/data, JAX agents, and TensorFlow/Sonnet agents.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill339Updated Sep 3, 2026

See all skills by VectorSpaceLab

Report a problem with this skill