Routes 3D ResNets PyTorch video action-recognition workflows across training, inference, and data preparation.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
VectorSpaceLab/AREX-Skill3392026년 9월 3일 업데이트
무료GitHub에 공개
fast-reid
Operate FastReID setup, configuration, datasets, model/inference APIs, training/evaluation workflows, deployment/export paths, and extension projects for person and vehicle re-identification.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
설치 방법 보기포함된 파일(36)
- SKILL.md4.1 KB
- references/package-map.md5.5 KB
- references/repo-provenance.md1.8 KB
- references/repo-routing-metadata.json514 B
- references/troubleshooting.md6.0 KB
- scripts/check_fastreid_environment.py6.4 KB
- sub-skills/data-and-datasets/references/data-api.md8.8 KB
- sub-skills/data-and-datasets/references/dataset-formats.md6.8 KB
- sub-skills/data-and-datasets/references/troubleshooting.md7.0 KB
- sub-skills/data-and-datasets/scripts/validate_dataset_layout.py18.6 KB
- sub-skills/data-and-datasets/SKILL.md3.9 KB
- sub-skills/deployment-and-projects/references/deployment-workflows.md10.9 KB
- sub-skills/deployment-and-projects/references/project-extensions.md8.6 KB
- sub-skills/deployment-and-projects/references/troubleshooting.md8.5 KB
- sub-skills/deployment-and-projects/scripts/check_deployment_dependencies.py7.0 KB
- sub-skills/deployment-and-projects/scripts/project_import_probe.py11.9 KB
- sub-skills/deployment-and-projects/SKILL.md3.3 KB
- sub-skills/modeling-and-inference/references/inference-workflows.md5.5 KB
- sub-skills/modeling-and-inference/references/model-api.md6.2 KB
- sub-skills/modeling-and-inference/references/troubleshooting.md5.8 KB
- sub-skills/modeling-and-inference/scripts/feature_extraction_smoke.py10.5 KB
- sub-skills/modeling-and-inference/scripts/model_forward_smoke.py6.0 KB
- sub-skills/modeling-and-inference/SKILL.md2.8 KB
- sub-skills/setup-and-configuration/references/configuration.md5.0 KB
- sub-skills/setup-and-configuration/references/model-zoo-and-recipes.md4.6 KB
- sub-skills/setup-and-configuration/references/troubleshooting.md4.4 KB
- sub-skills/setup-and-configuration/scripts/config_merge_check.py6.6 KB
- sub-skills/setup-and-configuration/SKILL.md2.5 KB
- sub-skills/training-and-evaluation/references/evaluation.md5.2 KB
- sub-skills/training-and-evaluation/references/trainer-api.md8.7 KB
- sub-skills/training-and-evaluation/references/training-cli.md10.3 KB
- sub-skills/training-and-evaluation/references/troubleshooting.md7.2 KB
- sub-skills/training-and-evaluation/scripts/run_training_entrypoint.py7.6 KB
- sub-skills/training-and-evaluation/scripts/train_command_builder.py11.2 KB
- sub-skills/training-and-evaluation/scripts/training_cli_help_check.py3.9 KB
- sub-skills/training-and-evaluation/SKILL.md3.8 KB
SKILL.md(원문)
설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.
FastReID repo skill
Use this skill when a task involves FastReID (fastreid), person or vehicle
re-identification, Market1501/DukeMTMC/MSMT17/VeRi/VehicleID style datasets,
FastReID YAML configs, ReID feature extraction, rank/mAP/mINP evaluation,
training/eval command construction, ONNX/Caffe/TensorRT export, or FastReID
extension projects.
First checks
- Confirm FastReID is importable. This inspected commit is source-only, so use
a local checkout import path or a bundled script's
--repo-rootoption when no installed distribution exists. - Confirm the intended backend.
MODEL.DEVICEdefaults tocuda; CPU smoke checks must override it explicitly. - Confirm required data/checkpoints. Real train/eval/inference requires local datasets and often local checkpoints or pretrain weights.
- Run the safe root helper when setup or optional dependencies are unclear:
python scripts/check_fastreid_environment.py \
--repo-root <FASTREID_REPO> \
--config-file <CONFIG_YAML> \
--device cpu
The helper imports modules and merges config only; it does not train, evaluate, download, or write checkpoints.
Route by task
| Task signal | Read |
|---|---|
Source-only import, dependencies, Python version, get_cfg, _BASE_, YAML merge, opts, model-zoo recipe/config selection | sub-skills/setup-and-configuration/ |
FASTREID_DATASETS, built-in dataset layouts, custom dataset registration, dataloaders, transforms, samplers, data preflight | sub-skills/data-and-datasets/ |
build_model, backbones/heads/losses/meta-architectures, DefaultPredictor, feature tensors, CPU model smoke, rank import quirks | sub-skills/modeling-and-inference/ |
Training/eval command building, DefaultTrainer, distributed flags, checkpoints/resume, solvers, metrics, logs, custom loops | sub-skills/training-and-evaluation/ |
| ONNX/Caffe/TensorRT export/inference, optional deployment deps, FastRT, project-extension imports/config hooks | sub-skills/deployment-and-projects/ |
Shared references
references/package-map.md— compact module, registry, signature, dependency, and route map for this FastReID version.references/troubleshooting.md— cross-cutting install/import, config, backend, dataset, checkpoint, inference, deployment, and project-extension failures.references/repo-provenance.md— source commit, package version, evidence paths, and refresh baseline.references/repo-routing-metadata.json— structured metadata for managed repo-skills-router import.
Shared script
scripts/check_fastreid_environment.py— safe import/config/optional-backend probe. Use it before deeper workflow debugging or before deciding whether CUDA/ONNX/Caffe/TensorRT dependencies are available.
Operating boundaries
- This skill teaches FastReID operation; it does not reproduce model-zoo benchmarks by itself.
- Do not run long training/evaluation, network downloads, dataset acquisition, Caffe/TensorRT builds, or multi-node jobs without explicit user data, hardware, runtime, and budget approval.
- CPU checks can validate imports/config/model plumbing, but they do not prove CUDA training performance or TensorRT/Caffe runtime equivalence.
- Extension projects add registry entries after importing their packages or
config hooks. If a project config fails with unknown keys or missing registry
entries, route to
deployment-and-projectsbefore editing configs.
리뷰
아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.
댓글
로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.
아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.
같은 저장소의 스킬
3ddfa
무료Guide 3DDFA Python inference, geometry rendering, training/evaluation, and optional C++ ONNX workflows for 3D dense face alignment.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
VectorSpaceLab/AREX-Skill3392026년 9월 3일 업데이트
3ddfa-v2
무료Routes 3DDFA_V2 face-alignment setup, still-image demos, video tracking, and ONNX benchmarking workflows.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
VectorSpaceLab/AREX-Skill3392026년 9월 3일 업데이트
ab3dmot
무료Operate AB3DMOT 3D multi-object tracking workflows for KITTI and nuScenes data, tracking, evaluation, and visualization.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
VectorSpaceLab/AREX-Skill3392026년 9월 3일 업데이트
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.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
VectorSpaceLab/AREX-Skill3392026년 9월 3일 업데이트
acme
무료Route Acme reinforcement-learning framework tasks across core loops, replay/data, JAX agents, and TensorFlow/Sonnet agents.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
VectorSpaceLab/AREX-Skill3392026년 9월 3일 업데이트