协作头脑风暴
brainstorming
在实现前澄清需求、探索设计选择,并形成清晰的开发方向。
training-and-cli【开发与工程】Config-driven training, conversion, debugging, and packaging workflows for spaCy.
独立 GitHub 快照 · 自动分类 · 以原作者文档为准
Enable config-driven training, evaluation, packaging, and safe CLI/data-conversion workflows for spaCy without large training runs or downloads by default.
.spacy;references/training-and-cli.md for the command map and recommended order of operations.references/config-reference.md for [paths], [system], [nlp], [components], [corpora], [training], [initialize], and [pretraining].references/data-formats-and-conversion.md for .spacy, DocBin, Corpus, Example, and conversion recipes.references/troubleshooting.md when config, data, threshold, package, or registry errors need triage.scripts/config_smoke.py to run a safe config round-trip and validation smoke.scripts/convert_textcat_jsonl_to_docbin.py to convert tiny textcat JSONL into .spacy without model downloads.spacy project commands go to project-workflows.pipeline-components.Doc, DocBin, tokenizer, matcher, and visualization details go to documents-and-visualization.install-and-inspect.init config or init fill-configdebug configconvertdebug datatrainevaluatepackagevalidatedebug config fails, fix config or registry references first.debug config passes but debug data fails, inspect paths, corpus files, or annotations next.Never run full training by default; keep training smoke tests tiny and CPU-first unless the environment explicitly proves otherwise.