[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7FCCJXR1STNYV1D7MCMSCF7":3,"skill-same:01M7FCCJXR1STNYV1D7MCMSCF7":67},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":28,"commitSha":29,"skillDirectory":30,"forksCount":31,"repositoryHtmlUrl":32,"repositoryHomepageUrl":6,"ownerHtmlUrl":33,"ownerBlogUrl":6,"ownerDescription":34,"updatedAt":35,"install":36,"resources":48},"01M7FCCJXR1STNYV1D7MCMSCF7","classification",null,"Use Simple Transformers classification APIs for text, sentence-pair, multilabel, LayoutLM\u002Fdocument, multimodal, regression, ONNX, and reranking workflows.","lza6\u002FArmor-piercing-system-prompts","lza6","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F128880206?v=4","01M7EB3K7WPX5T28EJ2NVHE6EG","Chinese-tingfeng","01M7EB3K97E1KQ16QWKJMFNGX8",11,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fsimpletransformers\u002Fsub-skills\u002Fclassification\u002FSKILL.md",[],[],"2026-11-16 10:05:34",false,"---\nname: classification\ndescription: \"Use Simple Transformers classification APIs for text,\n  sentence-pair, multilabel, LayoutLM\u002Fdocument, multimodal, regression, ONNX,\n  and reranking workflows.\"\ndisable-model-invocation: true\nmetadata:\n  disco-role: operating\nlicense: Apache 2.0\n---\n\n# Simple Transformers Classification Sub-skill\n\nUse this sub-skill when the user wants discriminative classification or\ncross-encoder reranking with the public `simpletransformers.classification`\nAPIs. It is self-contained: do not send future agents to the source checkout,\nrepo docs, tests, or examples for routine usage.\n\n## Owns\n\n- Binary text classification with `ClassificationModel`.\n- Multiclass text classification with `ClassificationModel(num_labels=N)`.\n- Regression with `ClassificationModel(num_labels=1)` and\n  `ClassificationArgs(regression=True)` or `args={\"regression\": True}`.\n- Sentence-pair classification\u002Fregression using `text_a`, `text_b`, `labels`\n  DataFrames and `predict([[text_a, text_b], ...])`.\n- Multi-label text classification with `MultiLabelClassificationModel` and\n  multi-hot list labels.\n- LayoutLM\u002FLayoutLMv2 document-style classification with `text`, `labels`,\n  `x0`, `y0`, `x1`, `y1` list columns and normalized 0-1000 boxes.\n- Multimodal text+image classification with `MultiModalClassificationModel`.\n- Cross-encoder reranker-style use of `ClassificationModel.rerank()` and\n  sentence-pair `predict()` scores.\n- Classification model conversion helpers such as `convert_to_onnx()` and\n  shared CPU-safe args for output\u002Fcache handling.\n\n## Route elsewhere\n\n- Named entity recognition and extractive question answering:\n  [token-and-qa](..\u002Ftoken-and-qa\u002FSKILL.md).\n- T5, Seq2Seq, language modeling, language generation, and text-to-text\n  classification\u002Fquestion generation:\n  [generative-workflows](..\u002Fgenerative-workflows\u002FSKILL.md).\n- Dense retrieval, representation vectors, DPR, BEIR\u002FMSMARCO data ownership,\n  and retrieval-model reranking:\n  [retrieval-representation](..\u002Fretrieval-representation\u002FSKILL.md).\n- Package installation, shared `ModelArgs`, Streamlit viewer, CUDA policy, and\n  cross-task output\u002Fcache conventions belong to the root skill.\n\n## Read first\n\n1. [API reference](references\u002Fapi-reference.md) for constructor signatures,\n   methods, return shapes, supported model-type strings, and compatibility\n   notes.\n2. [Data formats](references\u002Fdata-formats.md) before preparing any DataFrame,\n   lazy-loading file, LayoutLM box table, multimodal table, or reranking data.\n3. [Workflows](references\u002Fworkflows.md) for train\u002Feval\u002Fpredict recipes that do\n   not require opening original examples.\n4. [Troubleshooting](references\u002Ftroubleshooting.md) when imports fail, data is\n   silently misread, model downloads start unexpectedly, ONNX conversion fails,\n   or multimodal\u002FLayoutLM inputs are rejected.\n\n## Safe defaults\n\nPrefer explicit CPU-safe construction unless the user asks for GPU execution:\n\n```python\nfrom simpletransformers.classification import ClassificationArgs, ClassificationModel\n\nargs = ClassificationArgs()\nargs.no_save = True                 # avoid writing checkpoints in smoke runs\nargs.overwrite_output_dir = True    # required if reusing an output_dir\nargs.reprocess_input_data = True    # avoid stale cached features while debugging\nargs.max_seq_length = 128\nargs.scheduler = \"constant_schedule\"  # small smoke runs often do not need warmup\n\nmodel = ClassificationModel(\n    \"roberta\",\n    \"roberta-base\",\n    args=args,\n    use_cuda=False,\n)\n```\n\nFor real training, let the user choose a writable `output_dir`, remove\n`no_save=True`, and consider `evaluate_during_training`, `save_best_model`, and\ncheckpoint retention settings from the root configuration reference.\n\n## Data validation helper\n\nBefore training or evaluating, validate tabular inputs with the bundled helper:\n\n```bash\npython scripts\u002Fvalidate_classification_data.py --help\npython scripts\u002Fvalidate_classification_data.py --task single --input train.csv\npython scripts\u002Fvalidate_classification_data.py --task sentence-pair --input pairs.csv\npython scripts\u002Fvalidate_classification_data.py --task multilabel --input multilabel.jsonl --num-labels 6\npython scripts\u002Fvalidate_classification_data.py --task layoutlm --input layout.jsonl\npython scripts\u002Fvalidate_classification_data.py --task multimodal --input mm.csv --image-root images --check-image-exists\n```\n\nThe script performs deterministic schema checks only. It does not import\nSimple Transformers, download models, train, evaluate, or contact the network.\nUse non-zero exits as data-preparation blockers.\n\n## Required decision points\n\n- **Single text vs sentence pair:** choose `text`\u002F`labels` or\n  `text_a`\u002F`text_b`\u002F`labels` explicitly. If a DataFrame contains both shapes,\n  Simple Transformers will prefer `text`\u002F`labels`; fix ambiguous columns first.\n- **Label semantics:** binary and multiclass labels are integer class ids;\n  regression labels are floats; multi-label labels are list-like multi-hot\n  vectors of only `0`\u002F`1` values.\n- **Class count:** pass `num_labels` for multiclass, multilabel, and regression\n  (`num_labels=1`). For custom string labels, set `labels_list`\u002F`labels_map`\n  where supported and verify the mapped predictions.\n- **Long documents:** `ClassificationArgs.sliding_window=True` avoids hard\n  truncation for long single-text classification; it is not implemented for\n  multi-label and cannot be combined with lazy loading.\n- **LayoutLM:** box lists must align with whitespace word count and contain\n  normalized integer coordinates in `[0, 1000]` with `x0 \u003C= x1` and `y0 \u003C= y1`.\n- **Multimodal:** the DataFrame path stores relative image filenames in\n  `images`; `image_path` points to the directory containing those images.\n  `MultiModalClassificationModel` only supports `model_type=\"bert\"`.\n- **Reranking:** `ClassificationModel.rerank()` is for cross-encoder scoring;\n  retrieval dataset construction and dense retrieval ownership are routed to\n  `retrieval-representation`.\n\n## Minimal API map\n\n- Text\u002Fsentence-pair\u002Fregression\u002FLayoutLM:\n  `ClassificationModel`, `ClassificationArgs`.\n- Multi-label text: `MultiLabelClassificationModel`,\n  `MultiLabelClassificationArgs`.\n- Text+image multimodal: `MultiModalClassificationModel`,\n  `MultiModalClassificationArgs`.\n- Core methods: `train_model()`, `eval_model()`, `predict()`, `rerank()`,\n  `convert_to_onnx()`.\n- Custom classification losses recognized by `ClassificationModel` through\n  `args.loss_type`: `\"focal\"`, `\"dice\"`, or `\"tversky\"` with `loss_args`.\n\n## Verification status carried into this sub-skill\n\n- The public signatures and dataclass fields in this sub-skill were checked\n  against Simple Transformers 0.70.8 source and package-inspection evidence.\n- A safe native candidate named `test_custom_model_post_init` is relevant to\n  custom classification heads and Transformers compatibility.\n- Full native classification tests and examples download Hugging Face models\n  and run training; they are skipped by default unless the user explicitly\n  approves time, network\u002Fmodel cache, and compute budget.\n\n## Common next steps\n\n1. Identify which classification family the user wants.\n2. Validate the data schema using [data formats](references\u002Fdata-formats.md)\n   and the helper script.\n3. Instantiate the right model class with `use_cuda=False` for CPU-only or\n   smoke runs.\n4. Run the smallest possible `train_model()`\u002F`eval_model()`\u002F`predict()` loop,\n   then expand epochs, metrics, outputs, and backend settings only after schema\n   and import issues are resolved.\n5. If errors mention import aliases, MMBT, ONNX, cache reuse, pytrec, images,\n   or LayoutLM boxes, consult [troubleshooting](references\u002Ftroubleshooting.md)\n   before changing model architecture.\n","# Simple Transformers Classification Sub-skill\n\nUse this sub-skill when the user wants discriminative classification or\ncross-encoder reranking with the public `simpletransformers.classification`\nAPIs. It is self-contained: do not send future agents to the source checkout,\nrepo docs, tests, or examples for routine usage.\n\n## Owns\n\n- Binary text classification with `ClassificationModel`.\n- Multiclass text classification with `ClassificationModel(num_labels=N)`.\n- Regression with `ClassificationModel(num_labels=1)` and\n  `ClassificationArgs(regression=True)` or `args={\"regression\": True}`.\n- Sentence-pair classification\u002Fregression using `text_a`, `text_b`, `labels`\n  DataFrames and `predict([[text_a, text_b], ...])`.\n- Multi-label text classification with `MultiLabelClassificationModel` and\n  multi-hot list labels.\n- LayoutLM\u002FLayoutLMv2 document-style classification with `text`, `labels`,\n  `x0`, `y0`, `x1`, `y1` list columns and normalized 0-1000 boxes.\n- Multimodal text+image classification with `MultiModalClassificationModel`.\n- Cross-encoder reranker-style use of `ClassificationModel.rerank()` and\n  sentence-pair `predict()` scores.\n- Classification model conversion helpers such as `convert_to_onnx()` and\n  shared CPU-safe args for output\u002Fcache handling.\n\n## Route elsewhere\n\n- Named entity recognition and extractive question answering:\n  [token-and-qa](..\u002Ftoken-and-qa\u002FSKILL.md).\n- T5, Seq2Seq, language modeling, language generation, and text-to-text\n  classification\u002Fquestion generation:\n  [generative-workflows](..\u002Fgenerative-workflows\u002FSKILL.md).\n- Dense retrieval, representation vectors, DPR, BEIR\u002FMSMARCO data ownership,\n  and retrieval-model reranking:\n  [retrieval-representation](..\u002Fretrieval-representation\u002FSKILL.md).\n- Package installation, shared `ModelArgs`, Streamlit viewer, CUDA policy, and\n  cross-task output\u002Fcache conventions belong to the root skill.\n\n## Read first\n\n1. [API reference](references\u002Fapi-reference.md) for constructor signatures,\n   methods, return shapes, supported model-type strings, and compatibility\n   notes.\n2. [Data formats](references\u002Fdata-formats.md) before preparing any DataFrame,\n   lazy-loading file, LayoutLM box table, multimodal table, or reranking data.\n3. [Workflows](references\u002Fworkflows.md) for train\u002Feval\u002Fpredict recipes that do\n   not require opening original examples.\n4. [Troubleshooting](references\u002Ftroubleshooting.md) when imports fail, data is\n   silently misread, model downloads start unexpectedly, ONNX conversion fails,\n   or multimodal\u002FLayoutLM inputs are rejected.\n\n## Safe defaults\n\nPrefer explicit CPU-safe construction unless the user asks for GPU execution:\n\n```python\nfrom simpletransformers.classification import ClassificationArgs, ClassificationModel\n\nargs = ClassificationArgs()\nargs.no_save = True                 # avoid writing checkpoints in smoke runs\nargs.overwrite_output_dir = True    # required if reusing an output_dir\nargs.reprocess_input_data = True    # avoid stale cached features while debugging\nargs.max_seq_length = 128\nargs.scheduler = \"constant_schedule\"  # small smoke runs often do not need warmup\n\nmodel = ClassificationModel(\n    \"roberta\",\n    \"roberta-base\",\n    args=args,\n    use_cuda=False,\n)\n```\n\nFor real training, let the user choose a writable `output_dir`, remove\n`no_save=True`, and consider `evaluate_during_training`, `save_best_model`, and\ncheckpoint retention settings from the root configuration reference.\n\n## Data validation helper\n\nBefore training or evaluating, validate tabular inputs with the bundled helper:\n\n```bash\npython scripts\u002Fvalidate_classification_data.py --help\npython scripts\u002Fvalidate_classification_data.py --task single --input train.csv\npython scripts\u002Fvalidate_classification_data.py --task sentence-pair --input pairs.csv\npython scripts\u002Fvalidate_classification_data.py --task multilabel --input multilabel.jsonl --num-labels 6\npython scripts\u002Fvalidate_classification_data.py --task layoutlm --input layout.jsonl\npython scripts\u002Fvalidate_classification_data.py --task multimodal --input mm.csv --image-root images --check-image-exists\n```\n\nThe script performs deterministic schema checks only. It does not import\nSimple Transformers, download models, train, evaluate, or contact the network.\nUse non-zero exits as data-preparation blockers.\n\n## Required decision points\n\n- **Single text vs sentence pair:** choose `text`\u002F`labels` or\n  `text_a`\u002F`text_b`\u002F`labels` explicitly. If a DataFrame contains both shapes,\n  Simple Transformers will prefer `text`\u002F`labels`; fix ambiguous columns first.\n- **Label semantics:** binary and multiclass labels are integer class ids;\n  regression labels are floats; multi-label labels are list-like multi-hot\n  vectors of only `0`\u002F`1` values.\n- **Class count:** pass `num_labels` for multiclass, multilabel, and regression\n  (`num_labels=1`). For custom string labels, set `labels_list`\u002F`labels_map`\n  where supported and verify the mapped predictions.\n- **Long documents:** `ClassificationArgs.sliding_window=True` avoids hard\n  truncation for long single-text classification; it is not implemented for\n  multi-label and cannot be combined with lazy loading.\n- **LayoutLM:** box lists must align with whitespace word count and contain\n  normalized integer coordinates in `[0, 1000]` with `x0 \u003C= x1` and `y0 \u003C= y1`.\n- **Multimodal:** the DataFrame path stores relative image filenames in\n  `images`; `image_path` points to the directory containing those images.\n  `MultiModalClassificationModel` only supports `model_type=\"bert\"`.\n- **Reranking:** `ClassificationModel.rerank()` is for cross-encoder scoring;\n  retrieval dataset construction and dense retrieval ownership are routed to\n  `retrieval-representation`.\n\n## Minimal API map\n\n- Text\u002Fsentence-pair\u002Fregression\u002FLayoutLM:\n  `ClassificationModel`, `ClassificationArgs`.\n- Multi-label text: `MultiLabelClassificationModel`,\n  `MultiLabelClassificationArgs`.\n- Text+image multimodal: `MultiModalClassificationModel`,\n  `MultiModalClassificationArgs`.\n- Core methods: `train_model()`, `eval_model()`, `predict()`, `rerank()`,\n  `convert_to_onnx()`.\n- Custom classification losses recognized by `ClassificationModel` through\n  `args.loss_type`: `\"focal\"`, `\"dice\"`, or `\"tversky\"` with `loss_args`.\n\n## Verification status carried into this sub-skill\n\n- The public signatures and dataclass fields in this sub-skill were checked\n  against Simple Transformers 0.70.8 source and package-inspection evidence.\n- A safe native candidate named `test_custom_model_post_init` is relevant to\n  custom classification heads and Transformers compatibility.\n- Full native classification tests and examples download Hugging Face models\n  and run training; they are skipped by default unless the user explicitly\n  approves time, network\u002Fmodel cache, and compute budget.\n\n## Common next steps\n\n1. Identify which classification family the user wants.\n2. Validate the data schema using [data formats](references\u002Fdata-formats.md)\n   and the helper script.\n3. Instantiate the right model class with `use_cuda=False` for CPU-only or\n   smoke runs.\n4. Run the smallest possible `train_model()`\u002F`eval_model()`\u002F`predict()` loop,\n   then expand epochs, metrics, outputs, and backend settings only after schema\n   and import issues are resolved.\n5. If errors mention import aliases, MMBT, ONNX, cache reuse, pytrec, images,\n   or LayoutLM boxes, consult [troubleshooting](references\u002Ftroubleshooting.md)\n   before changing model architecture.\n",{"name":5,"license":24,"metadata":25,"description":7,"disable-model-invocation":27},"Apache 2.0",{"disco-role":26},"operating",true,"main","d69096e239f2a89aabc2cd078c47664a9d3b1e11","06-前端与设计\u002Fsimpletransformers\u002Fsub-skills\u002Fclassification",3,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts","https:\u002F\u002Fgithub.com\u002Flza6","个人vx：Tf00798\r\n个人✈：@TFGY999","2026-10-08 14:29:22",{"source":8,"skillName":5,"commands":37,"prompt":47,"zipAvailable":27},[38,41,44],{"tool":39,"command":40},"npx","npx skills add lza6\u002FArmor-piercing-system-prompts --skill classification",{"tool":42,"command":43},"bunx","bunx skills add lza6\u002FArmor-piercing-system-prompts --skill classification",{"tool":45,"command":46},"pnpm","pnpm dlx skills add lza6\u002FArmor-piercing-system-prompts --skill classification","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fsimpletransformers\u002Fsub-skills\u002Fclassification\u002FSKILL.md",[49,55,59,63],{"path":50,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":53,"sourceHtmlUrl":54},"references\u002Fapi-reference.md","REFERENCE","参考文档",4985,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fsimpletransformers\u002Fsub-skills\u002Fclassification\u002Freferences\u002Fapi-reference.md",{"path":56,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":57,"sourceHtmlUrl":58},"references\u002Fdata-formats.md",4450,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fsimpletransformers\u002Fsub-skills\u002Fclassification\u002Freferences\u002Fdata-formats.md",{"path":60,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":61,"sourceHtmlUrl":62},"references\u002Ftroubleshooting.md",3713,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fsimpletransformers\u002Fsub-skills\u002Fclassification\u002Freferences\u002Ftroubleshooting.md",{"path":64,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":65,"sourceHtmlUrl":66},"references\u002Fworkflows.md",3971,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F06-前端与设计\u002Fsimpletransformers\u002Fsub-skills\u002Fclassification\u002Freferences\u002Fworkflows.md",[68,76,81,88,95,102],{"skillNo":69,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":72,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M7FFG485KY83TD6HSC4YPCMC","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.","MIT","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator\u002FSKILL.md",[],[],{"skillNo":77,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":72,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":19},"01M7FFG0T6AZ0STEJPSB3CX84K","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator-antigravity-awesome-skills-main\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":72,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M7FFFY0QEM14SKWPDYGMXM2P","python-api","Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP\u002FmAP semantics, and source-style import diagnostics.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Fpython-api\u002FSKILL.md",[],[],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":72,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M7FFFVZQZ2GJ0C0TGP7JHD20","file-evaluation","Folder-based PASCAL VOC AP\u002FmAP evaluation from ground-truth and detection text files.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Ffile-evaluation\u002FSKILL.md",[],[],{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":72,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":101,"publishedAt":19},"01M7FFFSFNYSE77SJ6PE1CK2DA","object-detection-metrics","Guides legacy Object-Detection-Metrics PASCAL VOC AP\u002FmAP evaluation from detection text folders and direct Python API classes.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002FSKILL.md",[],[],{"skillNo":103,"name":104,"nameZh":6,"description":105,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":72,"licenseFileUrl":6,"sourceHtmlUrl":106,"categories":107,"keywords":108,"publishedAt":19},"01M7FFFP0GCK1RVXBCPQ8APBMT","nx-workspace-patterns","Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fnx-workspace-patterns\u002FSKILL.md",[],[]]