[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-tao-analyze-gaps-vlm-bcq":3,"mdc--cqkb7j-key":34,"related-repo-nvidia-tao-analyze-gaps-vlm-bcq":874,"related-org-nvidia-tao-analyze-gaps-vlm-bcq":978},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":23,"repoUrl":24,"updatedAt":25,"license":26,"forks":27,"topics":28,"repo":29,"sourceUrl":32,"mdContent":33},"tao-analyze-gaps-vlm-bcq","analyze VLM binary classification gaps","Extract false-positive and false-negative gaps from VLM binary-classification-question (BCQ, yes\u002Fno) predictions. Use when the user asks to \"analyze VLM BCQ gaps\", \"extract VLM false positives and false negatives\", or identify failure cases from a predictions JSON for DEFT root-cause analysis on a binary-classification VLM workflow.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,17,20],{"name":13,"slug":14,"type":15},"Data Analysis","data-analysis","tag",{"name":9,"slug":8,"type":15},{"name":18,"slug":19,"type":15},"Computer Vision","computer-vision",{"name":21,"slug":22,"type":15},"Debugging","debugging",2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-14T05:26:13.851319","Apache-2.0",281,[],{"repoUrl":24,"stars":23,"forks":27,"topics":30,"description":31},[],"AI agent skills published by NVIDIA","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Ftao-analyze-gaps-vlm-bcq","---\nname: tao-analyze-gaps-vlm-bcq\ndescription: Extract false-positive and false-negative gaps from VLM binary-classification-question (BCQ, yes\u002Fno) predictions.\n  Use when the user asks to \"analyze VLM BCQ gaps\", \"extract VLM false positives and false negatives\", or identify failure\n  cases from a predictions JSON for DEFT root-cause analysis on a binary-classification VLM workflow.\nlicense: Apache-2.0\ncompatibility: Requires docker + nvidia-container-toolkit.\nmetadata:\n  author: NVIDIA Corporation\n  version: \"0.1.0\"\nallowed-tools: Read Bash\ntags:\n- gap-analysis\n- rcca\n- vlm\n- evaluation\n- false-positive\n- false-negative\n---\n\n# VLM Binary Classification Gap Analysis\n\nReads a VLM predictions JSON, compares each model response against ground truth, and writes FP\u002FFN failure cases to a JSONL file with a summary report.\n\n## Purpose\n\nAfter running a VLM on a binary yes\u002Fno evaluation task, the predictions need to be compared against ground truth to identify failure cases. This skill produces a structured list of FP (false positive) and FN (false negative) samples that downstream RCCA stages (e.g., cosmos generation, root cause analysis) consume to drive a DEFT iteration.\n\n## Usage\n\nInvoke the `vlm_bcq` action inside the TAO Toolkit data services container with Hydra-style key=value overrides:\n\n```bash\ngap_analysis vlm_bcq \\\n  predictions_json=\u002Fpath\u002Fto\u002Fresults.json \\\n  results_dir=\u002Fpath\u002Fto\u002Foutput\u002Fgaps\n```\n\nInclude `videos_dir` when `video_id` values in the predictions are relative paths:\n\n```bash\ngap_analysis vlm_bcq \\\n  predictions_json=\u002Fpath\u002Fto\u002Fresults.json \\\n  results_dir=\u002Fpath\u002Fto\u002Foutput\u002Fgaps \\\n  videos_dir=\u002Fpath\u002Fto\u002Fvideos\u002Froot\n```\n\nAfter the run, surface the FP\u002FFN counts from `kpi_gaps_report.txt` and point downstream stages at `kpi_gaps.jsonl`.\n\n## Inputs\n\n- **predictions_json**: Path to predictions JSON file. Must be a JSON array where each item has `video_id`, `response`, and `gt` fields. `response` and `gt` are parsed with word-boundary matching — `'yes'` or `'no'` anywhere in the string is recognized. Samples where both or neither are present are skipped with a warning.\n- **videos_dir** (optional): Base directory for resolving relative `video_id` paths. If omitted, `video_id` values are used as absolute paths.\n\n**Predictions JSON format:**\n```json\n[\n  {\n    \"video_id\": \"\u002Fpath\u002Fto\u002Fvideo.mp4\",\n    \"response\": \"Yes, there is a collision.\",\n    \"gt\": \"B. No\",\n    \"question\": \"Is there a collision?\"\n  }\n]\n```\n\n## Outputs\n\n- **kpi_gaps.jsonl**: One JSON object per line for each FP\u002FFN case. Fields: `video_id` (absolute path), `error_type` (`FP` or `FN`), `question`, `ground_truth`, `response`.\n- **kpi_gaps_report.txt**: Human-readable table with total FP\u002FFN counts.\n\nIf no gaps are found, no files are written and a message is logged.\n\n## Key Parameters\n\n| Parameter | Required | Description |\n|-----------|----------|-------------|\n| predictions_json | Yes | Path to predictions JSON file |\n| results_dir | Yes | Output directory; created if it does not exist |\n| videos_dir | No | Base directory for resolving relative `video_id` paths |\n\n## Error Patterns\n\n| Error | Cause | Fix |\n|-------|-------|-----|\n| `FileNotFoundError` | `predictions_json` does not exist | Check the path |\n| `ValueError: must be a JSON array` | Predictions file is not a list | Wrap predictions in `[...]` |\n| `ValueError: missing 'gt'\u002F'response'\u002F'video_id'` | A prediction item is missing a required field | Inspect and fix the predictions JSON |\n| Samples silently skipped | `response` or `gt` contains both or neither 'yes'\u002F'no' | Check logs for warnings; inspect those samples |\n",{"data":35,"body":48},{"name":4,"description":6,"license":26,"compatibility":36,"metadata":37,"allowed-tools":40,"tags":41},"Requires docker + nvidia-container-toolkit.",{"author":38,"version":39},"NVIDIA Corporation","0.1.0","Read Bash",[42,43,44,45,46,47],"gap-analysis","rcca","vlm","evaluation","false-positive","false-negative",{"type":49,"children":50},"root",[51,60,66,73,78,84,98,155,176,230,251,257,349,357,551,557,627,632,638,729,735,868],{"type":52,"tag":53,"props":54,"children":56},"element","h1",{"id":55},"vlm-binary-classification-gap-analysis",[57],{"type":58,"value":59},"text","VLM Binary Classification Gap Analysis",{"type":52,"tag":61,"props":62,"children":63},"p",{},[64],{"type":58,"value":65},"Reads a VLM predictions JSON, compares each model response against ground truth, and writes FP\u002FFN failure cases to a JSONL file with a summary report.",{"type":52,"tag":67,"props":68,"children":70},"h2",{"id":69},"purpose",[71],{"type":58,"value":72},"Purpose",{"type":52,"tag":61,"props":74,"children":75},{},[76],{"type":58,"value":77},"After running a VLM on a binary yes\u002Fno evaluation task, the predictions need to be compared against ground truth to identify failure cases. This skill produces a structured list of FP (false positive) and FN (false negative) samples that downstream RCCA stages (e.g., cosmos generation, root cause analysis) consume to drive a DEFT iteration.",{"type":52,"tag":67,"props":79,"children":81},{"id":80},"usage",[82],{"type":58,"value":83},"Usage",{"type":52,"tag":61,"props":85,"children":86},{},[87,89,96],{"type":58,"value":88},"Invoke the ",{"type":52,"tag":90,"props":91,"children":93},"code",{"className":92},[],[94],{"type":58,"value":95},"vlm_bcq",{"type":58,"value":97}," action inside the TAO Toolkit data services container with Hydra-style key=value overrides:",{"type":52,"tag":99,"props":100,"children":105},"pre",{"className":101,"code":102,"language":103,"meta":104,"style":104},"language-bash shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","gap_analysis vlm_bcq \\\n  predictions_json=\u002Fpath\u002Fto\u002Fresults.json \\\n  results_dir=\u002Fpath\u002Fto\u002Foutput\u002Fgaps\n","bash","",[106],{"type":52,"tag":90,"props":107,"children":108},{"__ignoreMap":104},[109,133,146],{"type":52,"tag":110,"props":111,"children":114},"span",{"class":112,"line":113},"line",1,[115,121,127],{"type":52,"tag":110,"props":116,"children":118},{"style":117},"--shiki-light:#E2931D;--shiki-default:#FFCB6B;--shiki-dark:#FFCB6B",[119],{"type":58,"value":120},"gap_analysis",{"type":52,"tag":110,"props":122,"children":124},{"style":123},"--shiki-light:#91B859;--shiki-default:#C3E88D;--shiki-dark:#C3E88D",[125],{"type":58,"value":126}," vlm_bcq",{"type":52,"tag":110,"props":128,"children":130},{"style":129},"--shiki-light:#90A4AE;--shiki-default:#EEFFFF;--shiki-dark:#BABED8",[131],{"type":58,"value":132}," \\\n",{"type":52,"tag":110,"props":134,"children":136},{"class":112,"line":135},2,[137,142],{"type":52,"tag":110,"props":138,"children":139},{"style":123},[140],{"type":58,"value":141},"  predictions_json=\u002Fpath\u002Fto\u002Fresults.json",{"type":52,"tag":110,"props":143,"children":144},{"style":129},[145],{"type":58,"value":132},{"type":52,"tag":110,"props":147,"children":149},{"class":112,"line":148},3,[150],{"type":52,"tag":110,"props":151,"children":152},{"style":123},[153],{"type":58,"value":154},"  results_dir=\u002Fpath\u002Fto\u002Foutput\u002Fgaps\n",{"type":52,"tag":61,"props":156,"children":157},{},[158,160,166,168,174],{"type":58,"value":159},"Include ",{"type":52,"tag":90,"props":161,"children":163},{"className":162},[],[164],{"type":58,"value":165},"videos_dir",{"type":58,"value":167}," when ",{"type":52,"tag":90,"props":169,"children":171},{"className":170},[],[172],{"type":58,"value":173},"video_id",{"type":58,"value":175}," values in the predictions are relative paths:",{"type":52,"tag":99,"props":177,"children":179},{"className":101,"code":178,"language":103,"meta":104,"style":104},"gap_analysis vlm_bcq \\\n  predictions_json=\u002Fpath\u002Fto\u002Fresults.json \\\n  results_dir=\u002Fpath\u002Fto\u002Foutput\u002Fgaps \\\n  videos_dir=\u002Fpath\u002Fto\u002Fvideos\u002Froot\n",[180],{"type":52,"tag":90,"props":181,"children":182},{"__ignoreMap":104},[183,198,209,221],{"type":52,"tag":110,"props":184,"children":185},{"class":112,"line":113},[186,190,194],{"type":52,"tag":110,"props":187,"children":188},{"style":117},[189],{"type":58,"value":120},{"type":52,"tag":110,"props":191,"children":192},{"style":123},[193],{"type":58,"value":126},{"type":52,"tag":110,"props":195,"children":196},{"style":129},[197],{"type":58,"value":132},{"type":52,"tag":110,"props":199,"children":200},{"class":112,"line":135},[201,205],{"type":52,"tag":110,"props":202,"children":203},{"style":123},[204],{"type":58,"value":141},{"type":52,"tag":110,"props":206,"children":207},{"style":129},[208],{"type":58,"value":132},{"type":52,"tag":110,"props":210,"children":211},{"class":112,"line":148},[212,217],{"type":52,"tag":110,"props":213,"children":214},{"style":123},[215],{"type":58,"value":216},"  results_dir=\u002Fpath\u002Fto\u002Foutput\u002Fgaps",{"type":52,"tag":110,"props":218,"children":219},{"style":129},[220],{"type":58,"value":132},{"type":52,"tag":110,"props":222,"children":224},{"class":112,"line":223},4,[225],{"type":52,"tag":110,"props":226,"children":227},{"style":123},[228],{"type":58,"value":229},"  videos_dir=\u002Fpath\u002Fto\u002Fvideos\u002Froot\n",{"type":52,"tag":61,"props":231,"children":232},{},[233,235,241,243,249],{"type":58,"value":234},"After the run, surface the FP\u002FFN counts from ",{"type":52,"tag":90,"props":236,"children":238},{"className":237},[],[239],{"type":58,"value":240},"kpi_gaps_report.txt",{"type":58,"value":242}," and point downstream stages at ",{"type":52,"tag":90,"props":244,"children":246},{"className":245},[],[247],{"type":58,"value":248},"kpi_gaps.jsonl",{"type":58,"value":250},".",{"type":52,"tag":67,"props":252,"children":254},{"id":253},"inputs",[255],{"type":58,"value":256},"Inputs",{"type":52,"tag":258,"props":259,"children":260},"ul",{},[261,326],{"type":52,"tag":262,"props":263,"children":264},"li",{},[265,271,273,278,280,286,288,294,296,301,303,308,310,316,318,324],{"type":52,"tag":266,"props":267,"children":268},"strong",{},[269],{"type":58,"value":270},"predictions_json",{"type":58,"value":272},": Path to predictions JSON file. Must be a JSON array where each item has ",{"type":52,"tag":90,"props":274,"children":276},{"className":275},[],[277],{"type":58,"value":173},{"type":58,"value":279},", ",{"type":52,"tag":90,"props":281,"children":283},{"className":282},[],[284],{"type":58,"value":285},"response",{"type":58,"value":287},", and ",{"type":52,"tag":90,"props":289,"children":291},{"className":290},[],[292],{"type":58,"value":293},"gt",{"type":58,"value":295}," fields. ",{"type":52,"tag":90,"props":297,"children":299},{"className":298},[],[300],{"type":58,"value":285},{"type":58,"value":302}," and ",{"type":52,"tag":90,"props":304,"children":306},{"className":305},[],[307],{"type":58,"value":293},{"type":58,"value":309}," are parsed with word-boundary matching — ",{"type":52,"tag":90,"props":311,"children":313},{"className":312},[],[314],{"type":58,"value":315},"'yes'",{"type":58,"value":317}," or ",{"type":52,"tag":90,"props":319,"children":321},{"className":320},[],[322],{"type":58,"value":323},"'no'",{"type":58,"value":325}," anywhere in the string is recognized. Samples where both or neither are present are skipped with a warning.",{"type":52,"tag":262,"props":327,"children":328},{},[329,333,335,340,342,347],{"type":52,"tag":266,"props":330,"children":331},{},[332],{"type":58,"value":165},{"type":58,"value":334}," (optional): Base directory for resolving relative ",{"type":52,"tag":90,"props":336,"children":338},{"className":337},[],[339],{"type":58,"value":173},{"type":58,"value":341}," paths. If omitted, ",{"type":52,"tag":90,"props":343,"children":345},{"className":344},[],[346],{"type":58,"value":173},{"type":58,"value":348}," values are used as absolute paths.",{"type":52,"tag":61,"props":350,"children":351},{},[352],{"type":52,"tag":266,"props":353,"children":354},{},[355],{"type":58,"value":356},"Predictions JSON format:",{"type":52,"tag":99,"props":358,"children":362},{"className":359,"code":360,"language":361,"meta":104,"style":104},"language-json shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","[\n  {\n    \"video_id\": \"\u002Fpath\u002Fto\u002Fvideo.mp4\",\n    \"response\": \"Yes, there is a collision.\",\n    \"gt\": \"B. No\",\n    \"question\": \"Is there a collision?\"\n  }\n]\n","json",[363],{"type":52,"tag":90,"props":364,"children":365},{"__ignoreMap":104},[366,375,383,425,461,498,533,542],{"type":52,"tag":110,"props":367,"children":368},{"class":112,"line":113},[369],{"type":52,"tag":110,"props":370,"children":372},{"style":371},"--shiki-light:#39ADB5;--shiki-default:#89DDFF;--shiki-dark:#89DDFF",[373],{"type":58,"value":374},"[\n",{"type":52,"tag":110,"props":376,"children":377},{"class":112,"line":135},[378],{"type":52,"tag":110,"props":379,"children":380},{"style":371},[381],{"type":58,"value":382},"  {\n",{"type":52,"tag":110,"props":384,"children":385},{"class":112,"line":148},[386,391,396,401,406,411,416,420],{"type":52,"tag":110,"props":387,"children":388},{"style":371},[389],{"type":58,"value":390},"    \"",{"type":52,"tag":110,"props":392,"children":394},{"style":393},"--shiki-light:#9C3EDA;--shiki-default:#C792EA;--shiki-dark:#C792EA",[395],{"type":58,"value":173},{"type":52,"tag":110,"props":397,"children":398},{"style":371},[399],{"type":58,"value":400},"\"",{"type":52,"tag":110,"props":402,"children":403},{"style":371},[404],{"type":58,"value":405},":",{"type":52,"tag":110,"props":407,"children":408},{"style":371},[409],{"type":58,"value":410}," \"",{"type":52,"tag":110,"props":412,"children":413},{"style":123},[414],{"type":58,"value":415},"\u002Fpath\u002Fto\u002Fvideo.mp4",{"type":52,"tag":110,"props":417,"children":418},{"style":371},[419],{"type":58,"value":400},{"type":52,"tag":110,"props":421,"children":422},{"style":371},[423],{"type":58,"value":424},",\n",{"type":52,"tag":110,"props":426,"children":427},{"class":112,"line":223},[428,432,436,440,444,448,453,457],{"type":52,"tag":110,"props":429,"children":430},{"style":371},[431],{"type":58,"value":390},{"type":52,"tag":110,"props":433,"children":434},{"style":393},[435],{"type":58,"value":285},{"type":52,"tag":110,"props":437,"children":438},{"style":371},[439],{"type":58,"value":400},{"type":52,"tag":110,"props":441,"children":442},{"style":371},[443],{"type":58,"value":405},{"type":52,"tag":110,"props":445,"children":446},{"style":371},[447],{"type":58,"value":410},{"type":52,"tag":110,"props":449,"children":450},{"style":123},[451],{"type":58,"value":452},"Yes, there is a collision.",{"type":52,"tag":110,"props":454,"children":455},{"style":371},[456],{"type":58,"value":400},{"type":52,"tag":110,"props":458,"children":459},{"style":371},[460],{"type":58,"value":424},{"type":52,"tag":110,"props":462,"children":464},{"class":112,"line":463},5,[465,469,473,477,481,485,490,494],{"type":52,"tag":110,"props":466,"children":467},{"style":371},[468],{"type":58,"value":390},{"type":52,"tag":110,"props":470,"children":471},{"style":393},[472],{"type":58,"value":293},{"type":52,"tag":110,"props":474,"children":475},{"style":371},[476],{"type":58,"value":400},{"type":52,"tag":110,"props":478,"children":479},{"style":371},[480],{"type":58,"value":405},{"type":52,"tag":110,"props":482,"children":483},{"style":371},[484],{"type":58,"value":410},{"type":52,"tag":110,"props":486,"children":487},{"style":123},[488],{"type":58,"value":489},"B. No",{"type":52,"tag":110,"props":491,"children":492},{"style":371},[493],{"type":58,"value":400},{"type":52,"tag":110,"props":495,"children":496},{"style":371},[497],{"type":58,"value":424},{"type":52,"tag":110,"props":499,"children":501},{"class":112,"line":500},6,[502,506,511,515,519,523,528],{"type":52,"tag":110,"props":503,"children":504},{"style":371},[505],{"type":58,"value":390},{"type":52,"tag":110,"props":507,"children":508},{"style":393},[509],{"type":58,"value":510},"question",{"type":52,"tag":110,"props":512,"children":513},{"style":371},[514],{"type":58,"value":400},{"type":52,"tag":110,"props":516,"children":517},{"style":371},[518],{"type":58,"value":405},{"type":52,"tag":110,"props":520,"children":521},{"style":371},[522],{"type":58,"value":410},{"type":52,"tag":110,"props":524,"children":525},{"style":123},[526],{"type":58,"value":527},"Is there a collision?",{"type":52,"tag":110,"props":529,"children":530},{"style":371},[531],{"type":58,"value":532},"\"\n",{"type":52,"tag":110,"props":534,"children":536},{"class":112,"line":535},7,[537],{"type":52,"tag":110,"props":538,"children":539},{"style":371},[540],{"type":58,"value":541},"  }\n",{"type":52,"tag":110,"props":543,"children":545},{"class":112,"line":544},8,[546],{"type":52,"tag":110,"props":547,"children":548},{"style":371},[549],{"type":58,"value":550},"]\n",{"type":52,"tag":67,"props":552,"children":554},{"id":553},"outputs",[555],{"type":58,"value":556},"Outputs",{"type":52,"tag":258,"props":558,"children":559},{},[560,618],{"type":52,"tag":262,"props":561,"children":562},{},[563,567,569,574,576,582,584,590,591,597,599,604,605,611,612,617],{"type":52,"tag":266,"props":564,"children":565},{},[566],{"type":58,"value":248},{"type":58,"value":568},": One JSON object per line for each FP\u002FFN case. Fields: ",{"type":52,"tag":90,"props":570,"children":572},{"className":571},[],[573],{"type":58,"value":173},{"type":58,"value":575}," (absolute path), ",{"type":52,"tag":90,"props":577,"children":579},{"className":578},[],[580],{"type":58,"value":581},"error_type",{"type":58,"value":583}," (",{"type":52,"tag":90,"props":585,"children":587},{"className":586},[],[588],{"type":58,"value":589},"FP",{"type":58,"value":317},{"type":52,"tag":90,"props":592,"children":594},{"className":593},[],[595],{"type":58,"value":596},"FN",{"type":58,"value":598},"), ",{"type":52,"tag":90,"props":600,"children":602},{"className":601},[],[603],{"type":58,"value":510},{"type":58,"value":279},{"type":52,"tag":90,"props":606,"children":608},{"className":607},[],[609],{"type":58,"value":610},"ground_truth",{"type":58,"value":279},{"type":52,"tag":90,"props":613,"children":615},{"className":614},[],[616],{"type":58,"value":285},{"type":58,"value":250},{"type":52,"tag":262,"props":619,"children":620},{},[621,625],{"type":52,"tag":266,"props":622,"children":623},{},[624],{"type":58,"value":240},{"type":58,"value":626},": Human-readable table with total FP\u002FFN counts.",{"type":52,"tag":61,"props":628,"children":629},{},[630],{"type":58,"value":631},"If no gaps are found, no files are written and a message is logged.",{"type":52,"tag":67,"props":633,"children":635},{"id":634},"key-parameters",[636],{"type":58,"value":637},"Key Parameters",{"type":52,"tag":639,"props":640,"children":641},"table",{},[642,666],{"type":52,"tag":643,"props":644,"children":645},"thead",{},[646],{"type":52,"tag":647,"props":648,"children":649},"tr",{},[650,656,661],{"type":52,"tag":651,"props":652,"children":653},"th",{},[654],{"type":58,"value":655},"Parameter",{"type":52,"tag":651,"props":657,"children":658},{},[659],{"type":58,"value":660},"Required",{"type":52,"tag":651,"props":662,"children":663},{},[664],{"type":58,"value":665},"Description",{"type":52,"tag":667,"props":668,"children":669},"tbody",{},[670,688,705],{"type":52,"tag":647,"props":671,"children":672},{},[673,678,683],{"type":52,"tag":674,"props":675,"children":676},"td",{},[677],{"type":58,"value":270},{"type":52,"tag":674,"props":679,"children":680},{},[681],{"type":58,"value":682},"Yes",{"type":52,"tag":674,"props":684,"children":685},{},[686],{"type":58,"value":687},"Path to predictions JSON file",{"type":52,"tag":647,"props":689,"children":690},{},[691,696,700],{"type":52,"tag":674,"props":692,"children":693},{},[694],{"type":58,"value":695},"results_dir",{"type":52,"tag":674,"props":697,"children":698},{},[699],{"type":58,"value":682},{"type":52,"tag":674,"props":701,"children":702},{},[703],{"type":58,"value":704},"Output directory; created if it does not exist",{"type":52,"tag":647,"props":706,"children":707},{},[708,712,717],{"type":52,"tag":674,"props":709,"children":710},{},[711],{"type":58,"value":165},{"type":52,"tag":674,"props":713,"children":714},{},[715],{"type":58,"value":716},"No",{"type":52,"tag":674,"props":718,"children":719},{},[720,722,727],{"type":58,"value":721},"Base directory for resolving relative ",{"type":52,"tag":90,"props":723,"children":725},{"className":724},[],[726],{"type":58,"value":173},{"type":58,"value":728}," paths",{"type":52,"tag":67,"props":730,"children":732},{"id":731},"error-patterns",[733],{"type":58,"value":734},"Error Patterns",{"type":52,"tag":639,"props":736,"children":737},{},[738,759],{"type":52,"tag":643,"props":739,"children":740},{},[741],{"type":52,"tag":647,"props":742,"children":743},{},[744,749,754],{"type":52,"tag":651,"props":745,"children":746},{},[747],{"type":58,"value":748},"Error",{"type":52,"tag":651,"props":750,"children":751},{},[752],{"type":58,"value":753},"Cause",{"type":52,"tag":651,"props":755,"children":756},{},[757],{"type":58,"value":758},"Fix",{"type":52,"tag":667,"props":760,"children":761},{},[762,789,817,839],{"type":52,"tag":647,"props":763,"children":764},{},[765,774,784],{"type":52,"tag":674,"props":766,"children":767},{},[768],{"type":52,"tag":90,"props":769,"children":771},{"className":770},[],[772],{"type":58,"value":773},"FileNotFoundError",{"type":52,"tag":674,"props":775,"children":776},{},[777,782],{"type":52,"tag":90,"props":778,"children":780},{"className":779},[],[781],{"type":58,"value":270},{"type":58,"value":783}," does not exist",{"type":52,"tag":674,"props":785,"children":786},{},[787],{"type":58,"value":788},"Check the path",{"type":52,"tag":647,"props":790,"children":791},{},[792,801,806],{"type":52,"tag":674,"props":793,"children":794},{},[795],{"type":52,"tag":90,"props":796,"children":798},{"className":797},[],[799],{"type":58,"value":800},"ValueError: must be a JSON array",{"type":52,"tag":674,"props":802,"children":803},{},[804],{"type":58,"value":805},"Predictions file is not a list",{"type":52,"tag":674,"props":807,"children":808},{},[809,811],{"type":58,"value":810},"Wrap predictions in ",{"type":52,"tag":90,"props":812,"children":814},{"className":813},[],[815],{"type":58,"value":816},"[...]",{"type":52,"tag":647,"props":818,"children":819},{},[820,829,834],{"type":52,"tag":674,"props":821,"children":822},{},[823],{"type":52,"tag":90,"props":824,"children":826},{"className":825},[],[827],{"type":58,"value":828},"ValueError: missing 'gt'\u002F'response'\u002F'video_id'",{"type":52,"tag":674,"props":830,"children":831},{},[832],{"type":58,"value":833},"A prediction item is missing a required field",{"type":52,"tag":674,"props":835,"children":836},{},[837],{"type":58,"value":838},"Inspect and fix the predictions JSON",{"type":52,"tag":647,"props":840,"children":841},{},[842,847,863],{"type":52,"tag":674,"props":843,"children":844},{},[845],{"type":58,"value":846},"Samples silently skipped",{"type":52,"tag":674,"props":848,"children":849},{},[850,855,856,861],{"type":52,"tag":90,"props":851,"children":853},{"className":852},[],[854],{"type":58,"value":285},{"type":58,"value":317},{"type":52,"tag":90,"props":857,"children":859},{"className":858},[],[860],{"type":58,"value":293},{"type":58,"value":862}," contains both or neither 'yes'\u002F'no'",{"type":52,"tag":674,"props":864,"children":865},{},[866],{"type":58,"value":867},"Check logs for warnings; inspect those samples",{"type":52,"tag":869,"props":870,"children":871},"style",{},[872],{"type":58,"value":873},"html .light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html.light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"items":875,"total":977},[876,891,905,919,931,948,963],{"slug":877,"name":877,"fn":878,"description":879,"org":880,"tags":881,"stars":23,"repoUrl":24,"updatedAt":890},"accelerated-computing-cudf","accelerate data processing with cuDF","Official NVIDIA-authored guidance for NVIDIA cuDF GPU DataFrames, pandas acceleration, dask-cuDF, ETL, joins, groupby, CSV\u002FParquet I\u002FO, nullable semantics, and multi-GPU DataFrame workloads.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[882,883,886,887],{"name":13,"slug":14,"type":15},{"name":884,"slug":885,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":888,"slug":889,"type":15},"Performance","performance","2026-07-14T05:28:43.176466",{"slug":892,"name":892,"fn":893,"description":894,"org":895,"tags":896,"stars":23,"repoUrl":24,"updatedAt":904},"aiq-deploy","deploy and manage NVIDIA AI-Q infrastructure","Use when asked to install, deploy, run, validate, troubleshoot, or stop NVIDIA AI-Q Blueprint infrastructure.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[897,900,903],{"name":898,"slug":899,"type":15},"Deployment","deployment",{"name":901,"slug":902,"type":15},"Infrastructure","infrastructure",{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":906,"name":906,"fn":907,"description":908,"org":909,"tags":910,"stars":23,"repoUrl":24,"updatedAt":918},"aiq-research","conduct deep research with AI-Q","Use when asked to run deep research or AI-Q research through a reachable NVIDIA AI-Q Blueprint backend.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[911,914,915],{"name":912,"slug":913,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":916,"slug":917,"type":15},"Research","research","2026-07-14T05:28:06.816956",{"slug":920,"name":920,"fn":921,"description":922,"org":923,"tags":924,"stars":23,"repoUrl":24,"updatedAt":930},"amc-run-sample-calibration","run AMC sample dataset calibration","Run end-to-end calibration on the shipped sample dataset (sdg_08_2_sample_data_010926.zip) against a running AMC microservice. Use when user says 'test sample dataset', 'run sample calibration', 'verify AMC install', or 'launch and test'.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[925,926,927],{"name":13,"slug":14,"type":15},{"name":9,"slug":8,"type":15},{"name":928,"slug":929,"type":15},"Testing","testing","2026-07-17T05:29:03.913266",{"slug":932,"name":932,"fn":933,"description":934,"org":935,"tags":936,"stars":23,"repoUrl":24,"updatedAt":947},"amc-run-video-calibration","calibrate video datasets with AutoMagicCalib","Calibrate a new dataset from pre-recorded video files via the AutoMagicCalib REST API. Use when user has local MP4s and says 'calibrate my videos', 'run AMC on these videos', or similar. For RTSP\u002Flive streams, use amc-run-rtsp-calibration instead.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[937,940,943,944],{"name":938,"slug":939,"type":15},"Automation","automation",{"name":941,"slug":942,"type":15},"Imaging","imaging",{"name":9,"slug":8,"type":15},{"name":945,"slug":946,"type":15},"Video","video","2026-07-17T05:28:53.905004",{"slug":949,"name":949,"fn":950,"description":951,"org":952,"tags":953,"stars":23,"repoUrl":24,"updatedAt":962},"amc-setup-calibration-stack","deploy AutoMagicCalib microservice with Docker","Launch AutoMagicCalib microservice and web UI from NGC release images via Docker Compose. Use when user says 'deploy auto calibration', 'launch auto calibration', 'launch AMC', 'start MS+UI', or 'set up auto-magic-calib'. Requires NGC API key.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[954,955,958,959],{"name":898,"slug":899,"type":15},{"name":956,"slug":957,"type":15},"Docker","docker",{"name":9,"slug":8,"type":15},{"name":960,"slug":961,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":964,"name":964,"fn":965,"description":966,"org":967,"tags":968,"stars":23,"repoUrl":24,"updatedAt":976},"cudaq-guide","develop quantum applications with CUDA-Q","CUDA-Q onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[969,970,973],{"name":9,"slug":8,"type":15},{"name":971,"slug":972,"type":15},"Quantum Computing","quantum-computing",{"name":974,"slug":975,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305,{"items":979,"total":1128},[980,998,1014,1025,1037,1049,1062,1076,1087,1096,1110,1119],{"slug":981,"name":981,"fn":982,"description":983,"org":984,"tags":985,"stars":995,"repoUrl":996,"updatedAt":997},"nemoclaw-user-guide","retrieve NemoClaw documentation and configuration","Guides human users' AI agents to the NemoClaw docs MCP server and canonical Fern documentation in Markdown form. Use when users ask how to install, configure, operate, troubleshoot, secure, or learn NemoClaw with an AI coding assistant. Trigger keywords - nemoclaw docs, use nemoclaw with ai agent, nemoclaw mcp docs, nemoclaw install help, nemoclaw quickstart, nemoclaw markdown docs, llms.txt, agent skills.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[986,989,992],{"name":987,"slug":988,"type":15},"Documentation","documentation",{"name":990,"slug":991,"type":15},"MCP","mcp",{"name":993,"slug":994,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":999,"name":999,"fn":1000,"description":1001,"org":1002,"tags":1003,"stars":1011,"repoUrl":1012,"updatedAt":1013},"mcore-build-and-dependency","manage Megatron-LM development environments","Container-based dev environment setup and dependency management for Megatron-LM. Covers acquiring and launching the CI container, uv package management, and updating uv.lock.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1004,1007,1008],{"name":1005,"slug":1006,"type":15},"Containers","containers",{"name":898,"slug":899,"type":15},{"name":1009,"slug":1010,"type":15},"Python","python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":1015,"name":1015,"fn":1016,"description":1017,"org":1018,"tags":1019,"stars":1011,"repoUrl":1012,"updatedAt":1024},"mcore-bump-base-image","update NVIDIA PyTorch base images","Bump the NVIDIA PyTorch base image (`nvcr.io\u002Fnvidia\u002Fpytorch:YY.MM-py3`) used by Megatron-LM CI. Covers the two pin sites (GitHub CI in `docker\u002F.ngc_version.dev` and GitLab CI in `.gitlab\u002Fstages\u002F01.build.yml`), the post-bump CI loop (re-run functional tests, refresh golden values, mark broken tests), and the gotchas that bit PRs",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1020,1023],{"name":1021,"slug":1022,"type":15},"CI\u002FCD","ci-cd",{"name":898,"slug":899,"type":15},"2026-07-14T05:25:59.97109",{"slug":1026,"name":1026,"fn":1027,"description":1028,"org":1029,"tags":1030,"stars":1011,"repoUrl":1012,"updatedAt":1036},"mcore-cicd","manage CI\u002FCD pipelines for Megatron-LM","CI\u002FCD reference for Megatron-LM. Covers CI pipeline structure, PR scope labels, triggering internal GitLab CI (which force-pushes the current branch to a pull-request\u002FBRANCH ref — always dry-run and verify the destination first; never run against shared or protected branches), and CI failure investigation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1031,1032,1033],{"name":1021,"slug":1022,"type":15},{"name":898,"slug":899,"type":15},{"name":1034,"slug":1035,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":1038,"name":1038,"fn":1039,"description":1040,"org":1041,"tags":1042,"stars":1011,"repoUrl":1012,"updatedAt":1048},"mcore-create-issue","investigate CI failures and create issues","Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1043,1044,1045],{"name":21,"slug":22,"type":15},{"name":1034,"slug":1035,"type":15},{"name":1046,"slug":1047,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":1050,"name":1050,"fn":1051,"description":1052,"org":1053,"tags":1054,"stars":1011,"repoUrl":1012,"updatedAt":1061},"mcore-linting-and-formatting","lint and format Megatron-LM code","Linting and formatting for Megatron-LM. Covers running autoformat.sh, tools (ruff, black, isort, pylint, mypy), and code style rules.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1055,1058],{"name":1056,"slug":1057,"type":15},"Best Practices","best-practices",{"name":1059,"slug":1060,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":1063,"name":1063,"fn":1064,"description":1065,"org":1066,"tags":1067,"stars":1011,"repoUrl":1012,"updatedAt":1075},"mcore-migrate-gpt-to-hybrid","migrate Megatron-LM models to HybridModel","Migration guide for moving Megatron Core GPTModel checkpoints, model providers, training commands, and layer mappings to HybridModel.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1068,1071,1074],{"name":1069,"slug":1070,"type":15},"Machine Learning","machine-learning",{"name":1072,"slug":1073,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":1077,"name":1077,"fn":1078,"description":1079,"org":1080,"tags":1081,"stars":1011,"repoUrl":1012,"updatedAt":1086},"mcore-onboard-gb200-1node-tests","onboard functional tests for GB200","Onboard 1-node GitHub MR functional tests for GB200 from existing mr-scoped 2-node tests.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1082,1085],{"name":1083,"slug":1084,"type":15},"QA","qa",{"name":928,"slug":929,"type":15},"2026-07-14T05:25:53.673039",{"slug":1088,"name":1088,"fn":1089,"description":1090,"org":1091,"tags":1092,"stars":1011,"repoUrl":1012,"updatedAt":1095},"mcore-run-on-slurm","launch distributed training jobs on SLURM","How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-rank failure diagnosis.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1093,1094],{"name":898,"slug":899,"type":15},{"name":901,"slug":902,"type":15},"2026-07-14T05:25:49.362534",{"slug":1097,"name":1097,"fn":1098,"description":1099,"org":1100,"tags":1101,"stars":1011,"repoUrl":1012,"updatedAt":1109},"mcore-split-pr","split pull requests to reduce review load","Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1102,1105,1106],{"name":1103,"slug":1104,"type":15},"Code Review","code-review",{"name":1034,"slug":1035,"type":15},{"name":1107,"slug":1108,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":1111,"name":1111,"fn":1112,"description":1113,"org":1114,"tags":1115,"stars":1011,"repoUrl":1012,"updatedAt":1118},"mcore-testing","run and manage Megatron-LM tests","Test system for Megatron-LM. Covers test layout, recipe YAML structure, adding and running unit and functional tests, golden values, marker filters, and CI parity.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1116,1117],{"name":1083,"slug":1084,"type":15},{"name":928,"slug":929,"type":15},"2026-07-14T05:25:54.928983",{"slug":1120,"name":1120,"fn":1121,"description":1122,"org":1123,"tags":1124,"stars":1011,"repoUrl":1012,"updatedAt":1127},"nightly-sync","manage nightly main-to-dev sync workflows","Domain knowledge for the nightly main-to-dev sync workflow. Covers merge strategy, CI architecture, failure investigation, and known issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1125,1126],{"name":938,"slug":939,"type":15},{"name":1021,"slug":1022,"type":15},"2026-07-30T05:29:03.275638",496]