graph-data-and-datasets
Use CogDL graph data objects, custom datasets, built-in dataset selection, masks, batching, and no-download data validation.
- License 许可
- MIT
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- graph-data-and-datasets
- license
- MIT
- metadata
- { "disco-role": "operating" }
- description
- Use CogDL graph data objects, custom datasets, built-in dataset selection, masks, batching, and no-download data validation.
- disable-model-invocation
- true
02 资源文件 6
- 参考文档references/built-in-datasets.md7.2 KB
- 参考文档references/data-formats.md6.8 KB
- 参考文档references/graph-and-dataset-api.md9.6 KB
- 参考文档references/troubleshooting.md6.1 KB
- 脚本scripts/create_tiny_graph_dataset.py10.4 KB
- 脚本scripts/validate_graph_masks.py13.3 KB
03 SKILL.md 正文
CogDL Graph Data and Datasets
Use this sub-skill when the task is about CogDL data objects rather than model architecture or training orchestration.
Use when
- Creating or inspecting
cogdl.data.Graph,Adjacency,Dataset,NodeDataset,GraphDataset,generate_random_graph, orDataLoader. - Building a no-download custom node-classification or graph-classification fixture.
- Choosing a built-in dataset name while making download/cache/dependency effects explicit.
- Debugging
edge_index, CSR fields, masks, labels, node features, subgraphs,local_graph, self-loops, normalization, or batched graph objects.
Route elsewhere
- Training loops, wrapper matching, checkpoints, loggers, or
Trainerdetails:../training-wrappers-and-customization/SKILL.md. - High-level
experiment()calls, CLI flags, variants, or AutoML:../experiments-and-cli/SKILL.md. - Model, layer, sparse operator, PyG/Jittor/DGL, or CUDA acceleration details:
../models-layers-and-operators/SKILL.md. - Application pipelines such as OAG-BERT, recommendation, visualization, or
embedding pipelines:
../pipelines-and-applications/SKILL.md.
Operating checklist
- Prefer a self-contained
Graphor custom dataset artifact before loading a built-in dataset. Built-in datasets may download data and write caches when missing. - Validate node-classification fixtures before handing them to experiments or
wrappers. Use the bundled
scripts/validate_graph_masks.pyon a savedGraphorNodeDatasetartifact. - For quick no-download examples, use
scripts/create_tiny_graph_dataset.py --output-dir <directory>to create deterministic node and graph classification artifacts. - Keep
Graphschemas explicit: tensorx, tensory, COOedge_index, and boolean or index masks with node-count-compatible lengths. - For graph classification, save a list of local-node-indexed
Graphobjects and batch them withcogdl.data.DataLoader; if node features are absent, route the degree-feature training decision to the training-wrapper sub-skill.
Reference map
references/graph-and-dataset-api.md: verified public data APIs and graph methods.references/data-formats.md: COO/CSR formats, masks, batching, mutation semantics, and validation checklist.references/built-in-datasets.md: dataset families, registered names, and network/cache/dependency caveats.references/troubleshooting.md: common failures and recovery steps for graph data construction.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).