[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-nv-segment-ctmr":3,"mdc--uaawh1-key":31,"related-org-nvidia-nv-segment-ctmr":1919,"related-repo-nvidia-nv-segment-ctmr":2079},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":20,"repoUrl":21,"updatedAt":22,"license":23,"forks":24,"topics":25,"repo":26,"sourceUrl":29,"mdContent":30},"nv-segment-ctmr","run medical image segmentation on volumes","Used for running NV-Segment-CTMR on CT or MRI NIfTI volumes and recording label-map evidence. Not for clinical interpretation.",{"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],{"name":13,"slug":14,"type":15},"Healthcare","healthcare","tag",{"name":9,"slug":8,"type":15},{"name":18,"slug":19,"type":15},"Imaging","imaging",2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-14T05:29:55.619308","Apache-2.0",281,[],{"repoUrl":21,"stars":20,"forks":24,"topics":27,"description":28},[],"AI agent skills published by NVIDIA","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Fnv-segment-ctmr","---\nname: nv-segment-ctmr\ndescription: Used for running NV-Segment-CTMR on CT or MRI NIfTI volumes and recording label-map evidence. Not for clinical interpretation.\nlicense: Apache-2.0\nallowed-tools: Bash\nmetadata:\n  author: NVIDIA MedTech Team\n  tags:\n    - MedTech\n    - CT-MR\n    - segmentation\n---\n\n# NV-Segment-CTMR\n\n## Purpose\n- Used for running NV-Segment-CTMR on CT or MRI NIfTI volumes and recording label-map evidence. Not for clinical interpretation.\n- Use the wrapper exactly as documented; do not replace the upstream entrypoint with a handwritten implementation.\n- Manifest I\u002FO: inputs are `ct_or_mr_volume`; outputs are `label_map` and `result_json`.\n\n## Instructions\n- Read `skill_manifest.yaml` before changing arguments, side effects, or validation gates.\n- Run `scripts\u002Frun_ctmr.py` through the documented command below; keep outputs under a caller-provided run directory.\n- If a host agent exposes `run_script`, use `run_script(\"scripts\u002Frun_ctmr.py\", args=[...])`; otherwise run the Bash\u002FPython command shown below.\n- Check the emitted JSON and paired verifier guidance before treating the run as evidence.\n\n## Available Scripts\n| Script | Purpose | Arguments |\n|---|---|---|\n| `scripts\u002Frun_ctmr.py` | Primary entrypoint declared by skill_manifest.yaml. | `PATH_TO_IMAGE.nii.gz --output-dir OUT_DIR --modality CT_BODY [--label-prompts IDS]` |\n\n## Prerequisites\n- Runtime requirements: GPU\u002FCUDA when declared by the manifest; Python packages listed in `runtime.side_effects.pip_packages`.\n- Side effects: writes segmentation outputs under the caller's `--output-dir`, may cache model assets under `~\u002F.cache\u002Fhuggingface\u002F`, and may contact `https:\u002F\u002Fgithub.com` or `https:\u002F\u002Fhuggingface.co` during setup.\n- Run commands from the repository root unless an existing section below says otherwise.\n\n## Limitations\n- This is a thin wrapper. Inference, preprocessing, and postprocessing are delegated entirely to the upstream MONAI bundle under $NV_SEGMENT_CTMR_ROOT or the repo-local fallback at .workbench_data\u002Fupstreams\u002FNV-Segment-CTMR\u002FNV-Segment-CTMR.\n- The default wrapper path runs automatic \"segment everything\" inference for CT_BODY, MRI_BODY, or MRI_BRAIN. MRI_BRAIN inputs must already follow the upstream brain preprocessing requirements.\n- Label names are loaded from upstream configs when available. If a label dictionary is absent, the wrapper still records label IDs and marks only negative IDs as invalid.\n- No clinical, diagnostic, regulatory, or treatment-planning claims.\n- Not for clinical deployment, clinical interpretation, autonomous diagnosis, regulatory submission.\n\n## Troubleshooting\n| Error | Cause | Fix |\n|---|---|---|\n| Missing dependency or import error | Runtime package drift from `skill_manifest.yaml`. | Install the packages declared in the manifest or use the documented setup command. |\n| Empty or schema-invalid output | Wrong input path, unsupported modality, or upstream failure. | Re-run with a known fixture and inspect the wrapper JSON plus stderr. |\n| Validation gate failure | Output violated a declared engineering invariant. | Keep the failed evidence pack and use the gate message to repair inputs or wrapper code. |\n\nWraps the upstream\n[`NVIDIA-Medtech\u002FNV-Segment-CTMR`](https:\u002F\u002Fgithub.com\u002FNVIDIA-Medtech\u002FNV-Segment-CTMR\u002Ftree\u002Ff9f5f51b589e5dc9c23c453cf5138398e4084056\u002FNV-Segment-CTMR)\nCT\u002FMRI segmentation bundle. The wrapper does not reimplement VISTA3D\ninference. It shells out to the documented `python -m monai.bundle run`\nentry point, then inspects the produced NIfTI label map.\n\n\n## Exact Runnable Surface\n\nFor CT body segmentation user runs and benchmark answers, use this\nfresh-environment-safe repo-root command shape exactly:\n\n```bash\nexport NV_SEGMENT_CTMR_ROOT=\"${NV_SEGMENT_CTMR_ROOT:-$HOME\u002F.cache\u002Fnvidia-skills\u002Fupstreams\u002FNV-Segment-CTMR-f9f5f51\u002FNV-Segment-CTMR}\" && \\\npython -m pip install \"monai>=1.5,\u003C1.6\" \"numpy\u003C2\" nibabel scipy typer PyYAML fire huggingface_hub pytorch-ignite einops && \\\npython skills\u002Fnv-segment-ctmr\u002Fscripts\u002Frun_ctmr.py PATH_TO_IMAGE.nii.gz --modality CT_BODY --output-dir OUT_DIR\n```\n\nDo not invent `python -m nv_segment_ctmr`, `infer.py`, or `Medical AI Skills run` commands. `PATH_TO_IMAGE.nii.gz` must be the user's supplied input path.\nFor benchmark\u002Fuser run answers, the bash block is invalid if it includes\n`mkdir -p .workbench_data\u002Fupstreams`, `git clone`, `mkdir -p \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\"`,\n`hf download`, `mv \"$NV_SEGMENT_CTMR_ROOT\u002F...`, or any other command that\ncreates, downloads into, or moves files inside the shared upstream checkout.\n\n## Preconditions\n\nOne-time maintainer setup only; do not include these commands in user answers\nor benchmark commands. The benchmark environment already provides the\nrepo-local upstream cache and model files.\n\nIf `NV_SEGMENT_CTMR_ROOT` already names a local bundle checkout, the wrapper\nuses it and records its current commit in the result. Otherwise, clone the\nrecommended pinned default once:\n\n```bash\nif [ -z \"${NV_SEGMENT_CTMR_ROOT:-}\" ]; then\n  export NV_SEGMENT_CTMR_COMMIT=f9f5f51b589e5dc9c23c453cf5138398e4084056\n  export NV_SEGMENT_CTMR_CHECKOUT=\"$HOME\u002F.cache\u002Fnvidia-skills\u002Fupstreams\u002FNV-Segment-CTMR-f9f5f51\"\n  if [ ! -d \"$NV_SEGMENT_CTMR_CHECKOUT\u002F.git\" ]; then\n    git clone https:\u002F\u002Fgithub.com\u002FNVIDIA-Medtech\u002FNV-Segment-CTMR.git \"$NV_SEGMENT_CTMR_CHECKOUT\"\n    git -C \"$NV_SEGMENT_CTMR_CHECKOUT\" checkout --detach \"$NV_SEGMENT_CTMR_COMMIT\"\n  fi\n  export NV_SEGMENT_CTMR_ROOT=\"$NV_SEGMENT_CTMR_CHECKOUT\u002FNV-Segment-CTMR\"\nfi\npython -m pip install \"monai>=1.5,\u003C1.6\" \"numpy\u003C2\" nibabel scipy typer PyYAML fire huggingface_hub pytorch-ignite einops && \\\npython -c \"import monai, nibabel, numpy\"\n\nmkdir -p \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\"\ntest -e \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\u002Fmodel.pt\" || \\\n  hf download nvidia\u002FNV-Segment-CTMR \\\n    --revision 4fb8b4a6b2532be9f1c449a3726fe5440ab4213a \\\n    --local-dir \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\u002F\"\ntest -e \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\u002Fmodel.pt\" || \\\n  mv \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\u002Fvista3d_pretrained_model\u002Fmodel.pt\" \\\n    \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\u002Fmodel.pt\"\n```\n\nThe wrapper also searches `.workbench_data\u002Fupstreams\u002FNV-Segment-CTMR\u002FNV-Segment-CTMR`\nif `NV_SEGMENT_CTMR_ROOT` is unset or does not have the required bundle layout.\n\nFor agent-generated user run commands, use the command in Usage. Do not copy\nthe one-time Preconditions block into the answer: do not create or write under\n`$NV_SEGMENT_CTMR_ROOT`, do not run `hf download`, and do not move files in the\nshared upstream checkout during a benchmark or user run. Do not prepend\n`pip install -r \"$NV_SEGMENT_CTMR_ROOT\u002Frequirements.txt\"` in a Python 3.12\nenvironment; the upstream requirements pin NumPy 1.24.4, which does not build\ncleanly there. In a fresh Python environment, install the minimal compatible\nruntime shown above (`monai>=1.5,\u003C1.6`, `numpy\u003C2`, `nibabel`, `scipy`, `typer`,\n`PyYAML`, `fire`, `huggingface_hub`, `pytorch-ignite`, `einops`) before the\nwrapper. Cached models do not imply cached Python packages.\n\nRuntime needs an NVIDIA GPU with CUDA. The upstream bundle may import on\nCPU-only hosts, but this skill is declared as CUDA-required because the\npublished workflow is a 3D CT\u002FMRI foundation model inference path.\n\n## Usage\n\nFrom Medical AI Skills repo root:\n\n```bash\nexport NV_SEGMENT_CTMR_ROOT=\"${NV_SEGMENT_CTMR_ROOT:-$HOME\u002F.cache\u002Fnvidia-skills\u002Fupstreams\u002FNV-Segment-CTMR-f9f5f51\u002FNV-Segment-CTMR}\" && \\\npython -m pip install \"monai>=1.5,\u003C1.6\" \"numpy\u003C2\" nibabel scipy typer PyYAML fire huggingface_hub pytorch-ignite einops && \\\npython skills\u002Fnv-segment-ctmr\u002Fscripts\u002Frun_ctmr.py PATH_TO_IMAGE.nii.gz \\\n  --modality CT_BODY \\\n  --output-dir runs\u002Fnv_segment_ctmr_demo\n```\n\nReplace `PATH_TO_IMAGE.nii.gz` with the user's actual input path. Do not copy\nthe example fixture path into a user run. If the user provides an explicit\ninput path under `runs\u002F`, that path must be the first positional argument to\n`scripts\u002Frun_ctmr.py`.\n\nSupported automatic segmentation modalities are `CT_BODY`, `MRI_BODY`, and\n`MRI_BRAIN`. For `MRI_BRAIN`, the upstream README requires brain-specific\npreprocessing before bundle inference; pass an already preprocessed image to\nthis wrapper.\n\nPass `--label-prompts \"3,14\"` to request specific upstream class IDs instead\nof only the modality-level \"segment everything\" set. The evidence output\nrecords input geometry, output mask path, observed label IDs, unexpected\nlabels, per-class voxel counts, per-class physical volumes from the mask\nheader spacing, runtime, upstream command, model inventory, and geometry\nchecks.\n\nPass `--ground-truth PATH` to record a reference label-map path under\n`input.ground_truth_path`. The skill does not compute Dice; that is the\npaired verifier's job.\n\nAnatomy plausibility and optional per-class Dice\u002FIoU against the recorded\nground truth can be checked by `verifiers\u002Fct_segmentation_quality_v1` for\nCT-body outputs.\n\nNot for clinical interpretation, production deployment, autonomous diagnosis,\nor regulatory submission.\n",{"data":32,"body":40},{"name":4,"description":6,"license":23,"allowed-tools":33,"metadata":34},"Bash",{"author":35,"tags":36},"NVIDIA MedTech Team",[37,38,39],"MedTech","CT-MR","segmentation",{"type":41,"children":42},"root",[43,51,58,102,108,163,169,226,232,289,295,323,329,416,445,451,456,715,789,795,800,812,1444,1464,1560,1565,1571,1576,1801,1827,1862,1875,1895,1908,1913],{"type":44,"tag":45,"props":46,"children":47},"element","h1",{"id":4},[48],{"type":49,"value":50},"text","NV-Segment-CTMR",{"type":44,"tag":52,"props":53,"children":55},"h2",{"id":54},"purpose",[56],{"type":49,"value":57},"Purpose",{"type":44,"tag":59,"props":60,"children":61},"ul",{},[62,67,72],{"type":44,"tag":63,"props":64,"children":65},"li",{},[66],{"type":49,"value":6},{"type":44,"tag":63,"props":68,"children":69},{},[70],{"type":49,"value":71},"Use the wrapper exactly as documented; do not replace the upstream entrypoint with a handwritten implementation.",{"type":44,"tag":63,"props":73,"children":74},{},[75,77,84,86,92,94,100],{"type":49,"value":76},"Manifest I\u002FO: inputs are ",{"type":44,"tag":78,"props":79,"children":81},"code",{"className":80},[],[82],{"type":49,"value":83},"ct_or_mr_volume",{"type":49,"value":85},"; outputs are ",{"type":44,"tag":78,"props":87,"children":89},{"className":88},[],[90],{"type":49,"value":91},"label_map",{"type":49,"value":93}," and ",{"type":44,"tag":78,"props":95,"children":97},{"className":96},[],[98],{"type":49,"value":99},"result_json",{"type":49,"value":101},".",{"type":44,"tag":52,"props":103,"children":105},{"id":104},"instructions",[106],{"type":49,"value":107},"Instructions",{"type":44,"tag":59,"props":109,"children":110},{},[111,124,137,158],{"type":44,"tag":63,"props":112,"children":113},{},[114,116,122],{"type":49,"value":115},"Read ",{"type":44,"tag":78,"props":117,"children":119},{"className":118},[],[120],{"type":49,"value":121},"skill_manifest.yaml",{"type":49,"value":123}," before changing arguments, side effects, or validation gates.",{"type":44,"tag":63,"props":125,"children":126},{},[127,129,135],{"type":49,"value":128},"Run ",{"type":44,"tag":78,"props":130,"children":132},{"className":131},[],[133],{"type":49,"value":134},"scripts\u002Frun_ctmr.py",{"type":49,"value":136}," through the documented command below; keep outputs under a caller-provided run directory.",{"type":44,"tag":63,"props":138,"children":139},{},[140,142,148,150,156],{"type":49,"value":141},"If a host agent exposes ",{"type":44,"tag":78,"props":143,"children":145},{"className":144},[],[146],{"type":49,"value":147},"run_script",{"type":49,"value":149},", use ",{"type":44,"tag":78,"props":151,"children":153},{"className":152},[],[154],{"type":49,"value":155},"run_script(\"scripts\u002Frun_ctmr.py\", args=[...])",{"type":49,"value":157},"; otherwise run the Bash\u002FPython command shown below.",{"type":44,"tag":63,"props":159,"children":160},{},[161],{"type":49,"value":162},"Check the emitted JSON and paired verifier guidance before treating the run as evidence.",{"type":44,"tag":52,"props":164,"children":166},{"id":165},"available-scripts",[167],{"type":49,"value":168},"Available Scripts",{"type":44,"tag":170,"props":171,"children":172},"table",{},[173,196],{"type":44,"tag":174,"props":175,"children":176},"thead",{},[177],{"type":44,"tag":178,"props":179,"children":180},"tr",{},[181,187,191],{"type":44,"tag":182,"props":183,"children":184},"th",{},[185],{"type":49,"value":186},"Script",{"type":44,"tag":182,"props":188,"children":189},{},[190],{"type":49,"value":57},{"type":44,"tag":182,"props":192,"children":193},{},[194],{"type":49,"value":195},"Arguments",{"type":44,"tag":197,"props":198,"children":199},"tbody",{},[200],{"type":44,"tag":178,"props":201,"children":202},{},[203,212,217],{"type":44,"tag":204,"props":205,"children":206},"td",{},[207],{"type":44,"tag":78,"props":208,"children":210},{"className":209},[],[211],{"type":49,"value":134},{"type":44,"tag":204,"props":213,"children":214},{},[215],{"type":49,"value":216},"Primary entrypoint declared by skill_manifest.yaml.",{"type":44,"tag":204,"props":218,"children":219},{},[220],{"type":44,"tag":78,"props":221,"children":223},{"className":222},[],[224],{"type":49,"value":225},"PATH_TO_IMAGE.nii.gz --output-dir OUT_DIR --modality CT_BODY [--label-prompts IDS]",{"type":44,"tag":52,"props":227,"children":229},{"id":228},"prerequisites",[230],{"type":49,"value":231},"Prerequisites",{"type":44,"tag":59,"props":233,"children":234},{},[235,247,284],{"type":44,"tag":63,"props":236,"children":237},{},[238,240,246],{"type":49,"value":239},"Runtime requirements: GPU\u002FCUDA when declared by the manifest; Python packages listed in ",{"type":44,"tag":78,"props":241,"children":243},{"className":242},[],[244],{"type":49,"value":245},"runtime.side_effects.pip_packages",{"type":49,"value":101},{"type":44,"tag":63,"props":248,"children":249},{},[250,252,258,260,266,268,274,276,282],{"type":49,"value":251},"Side effects: writes segmentation outputs under the caller's ",{"type":44,"tag":78,"props":253,"children":255},{"className":254},[],[256],{"type":49,"value":257},"--output-dir",{"type":49,"value":259},", may cache model assets under ",{"type":44,"tag":78,"props":261,"children":263},{"className":262},[],[264],{"type":49,"value":265},"~\u002F.cache\u002Fhuggingface\u002F",{"type":49,"value":267},", and may contact ",{"type":44,"tag":78,"props":269,"children":271},{"className":270},[],[272],{"type":49,"value":273},"https:\u002F\u002Fgithub.com",{"type":49,"value":275}," or ",{"type":44,"tag":78,"props":277,"children":279},{"className":278},[],[280],{"type":49,"value":281},"https:\u002F\u002Fhuggingface.co",{"type":49,"value":283}," during setup.",{"type":44,"tag":63,"props":285,"children":286},{},[287],{"type":49,"value":288},"Run commands from the repository root unless an existing section below says otherwise.",{"type":44,"tag":52,"props":290,"children":292},{"id":291},"limitations",[293],{"type":49,"value":294},"Limitations",{"type":44,"tag":59,"props":296,"children":297},{},[298,303,308,313,318],{"type":44,"tag":63,"props":299,"children":300},{},[301],{"type":49,"value":302},"This is a thin wrapper. Inference, preprocessing, and postprocessing are delegated entirely to the upstream MONAI bundle under $NV_SEGMENT_CTMR_ROOT or the repo-local fallback at .workbench_data\u002Fupstreams\u002FNV-Segment-CTMR\u002FNV-Segment-CTMR.",{"type":44,"tag":63,"props":304,"children":305},{},[306],{"type":49,"value":307},"The default wrapper path runs automatic \"segment everything\" inference for CT_BODY, MRI_BODY, or MRI_BRAIN. MRI_BRAIN inputs must already follow the upstream brain preprocessing requirements.",{"type":44,"tag":63,"props":309,"children":310},{},[311],{"type":49,"value":312},"Label names are loaded from upstream configs when available. If a label dictionary is absent, the wrapper still records label IDs and marks only negative IDs as invalid.",{"type":44,"tag":63,"props":314,"children":315},{},[316],{"type":49,"value":317},"No clinical, diagnostic, regulatory, or treatment-planning claims.",{"type":44,"tag":63,"props":319,"children":320},{},[321],{"type":49,"value":322},"Not for clinical deployment, clinical interpretation, autonomous diagnosis, regulatory submission.",{"type":44,"tag":52,"props":324,"children":326},{"id":325},"troubleshooting",[327],{"type":49,"value":328},"Troubleshooting",{"type":44,"tag":170,"props":330,"children":331},{},[332,353],{"type":44,"tag":174,"props":333,"children":334},{},[335],{"type":44,"tag":178,"props":336,"children":337},{},[338,343,348],{"type":44,"tag":182,"props":339,"children":340},{},[341],{"type":49,"value":342},"Error",{"type":44,"tag":182,"props":344,"children":345},{},[346],{"type":49,"value":347},"Cause",{"type":44,"tag":182,"props":349,"children":350},{},[351],{"type":49,"value":352},"Fix",{"type":44,"tag":197,"props":354,"children":355},{},[356,380,398],{"type":44,"tag":178,"props":357,"children":358},{},[359,364,375],{"type":44,"tag":204,"props":360,"children":361},{},[362],{"type":49,"value":363},"Missing dependency or import error",{"type":44,"tag":204,"props":365,"children":366},{},[367,369,374],{"type":49,"value":368},"Runtime package drift from ",{"type":44,"tag":78,"props":370,"children":372},{"className":371},[],[373],{"type":49,"value":121},{"type":49,"value":101},{"type":44,"tag":204,"props":376,"children":377},{},[378],{"type":49,"value":379},"Install the packages declared in the manifest or use the documented setup command.",{"type":44,"tag":178,"props":381,"children":382},{},[383,388,393],{"type":44,"tag":204,"props":384,"children":385},{},[386],{"type":49,"value":387},"Empty or schema-invalid output",{"type":44,"tag":204,"props":389,"children":390},{},[391],{"type":49,"value":392},"Wrong input path, unsupported modality, or upstream failure.",{"type":44,"tag":204,"props":394,"children":395},{},[396],{"type":49,"value":397},"Re-run with a known fixture and inspect the wrapper JSON plus stderr.",{"type":44,"tag":178,"props":399,"children":400},{},[401,406,411],{"type":44,"tag":204,"props":402,"children":403},{},[404],{"type":49,"value":405},"Validation gate failure",{"type":44,"tag":204,"props":407,"children":408},{},[409],{"type":49,"value":410},"Output violated a declared engineering invariant.",{"type":44,"tag":204,"props":412,"children":413},{},[414],{"type":49,"value":415},"Keep the failed evidence pack and use the gate message to repair inputs or wrapper code.",{"type":44,"tag":417,"props":418,"children":419},"p",{},[420,422,435,437,443],{"type":49,"value":421},"Wraps the upstream\n",{"type":44,"tag":423,"props":424,"children":428},"a",{"href":425,"rel":426},"https:\u002F\u002Fgithub.com\u002FNVIDIA-Medtech\u002FNV-Segment-CTMR\u002Ftree\u002Ff9f5f51b589e5dc9c23c453cf5138398e4084056\u002FNV-Segment-CTMR",[427],"nofollow",[429],{"type":44,"tag":78,"props":430,"children":432},{"className":431},[],[433],{"type":49,"value":434},"NVIDIA-Medtech\u002FNV-Segment-CTMR",{"type":49,"value":436},"\nCT\u002FMRI segmentation bundle. The wrapper does not reimplement VISTA3D\ninference. It shells out to the documented ",{"type":44,"tag":78,"props":438,"children":440},{"className":439},[],[441],{"type":49,"value":442},"python -m monai.bundle run",{"type":49,"value":444},"\nentry point, then inspects the produced NIfTI label map.",{"type":44,"tag":52,"props":446,"children":448},{"id":447},"exact-runnable-surface",[449],{"type":49,"value":450},"Exact Runnable Surface",{"type":44,"tag":417,"props":452,"children":453},{},[454],{"type":49,"value":455},"For CT body segmentation user runs and benchmark answers, use this\nfresh-environment-safe repo-root command shape exactly:",{"type":44,"tag":457,"props":458,"children":463},"pre",{"className":459,"code":460,"language":461,"meta":462,"style":462},"language-bash shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","export NV_SEGMENT_CTMR_ROOT=\"${NV_SEGMENT_CTMR_ROOT:-$HOME\u002F.cache\u002Fnvidia-skills\u002Fupstreams\u002FNV-Segment-CTMR-f9f5f51\u002FNV-Segment-CTMR}\" && \\\npython -m pip install \"monai>=1.5,\u003C1.6\" \"numpy\u003C2\" nibabel scipy typer PyYAML fire huggingface_hub pytorch-ignite einops && \\\npython skills\u002Fnv-segment-ctmr\u002Fscripts\u002Frun_ctmr.py PATH_TO_IMAGE.nii.gz --modality CT_BODY --output-dir OUT_DIR\n","bash","",[464],{"type":44,"tag":78,"props":465,"children":466},{"__ignoreMap":462},[467,576,677],{"type":44,"tag":468,"props":469,"children":472},"span",{"class":470,"line":471},"line",1,[473,479,485,491,496,501,506,511,516,521,526,530,535,539,544,548,553,557,561,566,571],{"type":44,"tag":468,"props":474,"children":476},{"style":475},"--shiki-light:#9C3EDA;--shiki-default:#C792EA;--shiki-dark:#C792EA",[477],{"type":49,"value":478},"export",{"type":44,"tag":468,"props":480,"children":482},{"style":481},"--shiki-light:#90A4AE;--shiki-default:#EEFFFF;--shiki-dark:#BABED8",[483],{"type":49,"value":484}," NV_SEGMENT_CTMR_ROOT",{"type":44,"tag":468,"props":486,"children":488},{"style":487},"--shiki-light:#39ADB5;--shiki-default:#89DDFF;--shiki-dark:#89DDFF",[489],{"type":49,"value":490},"=",{"type":44,"tag":468,"props":492,"children":493},{"style":487},[494],{"type":49,"value":495},"\"${",{"type":44,"tag":468,"props":497,"children":498},{"style":481},[499],{"type":49,"value":500},"NV_SEGMENT_CTMR_ROOT",{"type":44,"tag":468,"props":502,"children":503},{"style":487},[504],{"type":49,"value":505},":-",{"type":44,"tag":468,"props":507,"children":508},{"style":481},[509],{"type":49,"value":510},"$HOME",{"type":44,"tag":468,"props":512,"children":513},{"style":487},[514],{"type":49,"value":515},"\u002F",{"type":44,"tag":468,"props":517,"children":519},{"style":518},"--shiki-light:#91B859;--shiki-default:#C3E88D;--shiki-dark:#C3E88D",[520],{"type":49,"value":101},{"type":44,"tag":468,"props":522,"children":523},{"style":481},[524],{"type":49,"value":525},"cache",{"type":44,"tag":468,"props":527,"children":528},{"style":487},[529],{"type":49,"value":515},{"type":44,"tag":468,"props":531,"children":532},{"style":481},[533],{"type":49,"value":534},"nvidia-skills",{"type":44,"tag":468,"props":536,"children":537},{"style":487},[538],{"type":49,"value":515},{"type":44,"tag":468,"props":540,"children":541},{"style":481},[542],{"type":49,"value":543},"upstreams",{"type":44,"tag":468,"props":545,"children":546},{"style":487},[547],{"type":49,"value":515},{"type":44,"tag":468,"props":549,"children":550},{"style":481},[551],{"type":49,"value":552},"NV-Segment-CTMR-f9f5f51",{"type":44,"tag":468,"props":554,"children":555},{"style":487},[556],{"type":49,"value":515},{"type":44,"tag":468,"props":558,"children":559},{"style":481},[560],{"type":49,"value":50},{"type":44,"tag":468,"props":562,"children":563},{"style":487},[564],{"type":49,"value":565},"}\"",{"type":44,"tag":468,"props":567,"children":568},{"style":487},[569],{"type":49,"value":570}," &&",{"type":44,"tag":468,"props":572,"children":573},{"style":481},[574],{"type":49,"value":575}," \\\n",{"type":44,"tag":468,"props":577,"children":579},{"class":470,"line":578},2,[580,586,591,596,601,606,611,616,620,625,629,634,639,644,649,654,659,664,669,673],{"type":44,"tag":468,"props":581,"children":583},{"style":582},"--shiki-light:#E2931D;--shiki-default:#FFCB6B;--shiki-dark:#FFCB6B",[584],{"type":49,"value":585},"python",{"type":44,"tag":468,"props":587,"children":588},{"style":518},[589],{"type":49,"value":590}," -m",{"type":44,"tag":468,"props":592,"children":593},{"style":518},[594],{"type":49,"value":595}," pip",{"type":44,"tag":468,"props":597,"children":598},{"style":518},[599],{"type":49,"value":600}," install",{"type":44,"tag":468,"props":602,"children":603},{"style":487},[604],{"type":49,"value":605}," \"",{"type":44,"tag":468,"props":607,"children":608},{"style":518},[609],{"type":49,"value":610},"monai>=1.5,\u003C1.6",{"type":44,"tag":468,"props":612,"children":613},{"style":487},[614],{"type":49,"value":615},"\"",{"type":44,"tag":468,"props":617,"children":618},{"style":487},[619],{"type":49,"value":605},{"type":44,"tag":468,"props":621,"children":622},{"style":518},[623],{"type":49,"value":624},"numpy\u003C2",{"type":44,"tag":468,"props":626,"children":627},{"style":487},[628],{"type":49,"value":615},{"type":44,"tag":468,"props":630,"children":631},{"style":518},[632],{"type":49,"value":633}," nibabel",{"type":44,"tag":468,"props":635,"children":636},{"style":518},[637],{"type":49,"value":638}," scipy",{"type":44,"tag":468,"props":640,"children":641},{"style":518},[642],{"type":49,"value":643}," typer",{"type":44,"tag":468,"props":645,"children":646},{"style":518},[647],{"type":49,"value":648}," PyYAML",{"type":44,"tag":468,"props":650,"children":651},{"style":518},[652],{"type":49,"value":653}," fire",{"type":44,"tag":468,"props":655,"children":656},{"style":518},[657],{"type":49,"value":658}," huggingface_hub",{"type":44,"tag":468,"props":660,"children":661},{"style":518},[662],{"type":49,"value":663}," pytorch-ignite",{"type":44,"tag":468,"props":665,"children":666},{"style":518},[667],{"type":49,"value":668}," einops",{"type":44,"tag":468,"props":670,"children":671},{"style":487},[672],{"type":49,"value":570},{"type":44,"tag":468,"props":674,"children":675},{"style":481},[676],{"type":49,"value":575},{"type":44,"tag":468,"props":678,"children":680},{"class":470,"line":679},3,[681,685,690,695,700,705,710],{"type":44,"tag":468,"props":682,"children":683},{"style":582},[684],{"type":49,"value":585},{"type":44,"tag":468,"props":686,"children":687},{"style":518},[688],{"type":49,"value":689}," skills\u002Fnv-segment-ctmr\u002Fscripts\u002Frun_ctmr.py",{"type":44,"tag":468,"props":691,"children":692},{"style":518},[693],{"type":49,"value":694}," PATH_TO_IMAGE.nii.gz",{"type":44,"tag":468,"props":696,"children":697},{"style":518},[698],{"type":49,"value":699}," --modality",{"type":44,"tag":468,"props":701,"children":702},{"style":518},[703],{"type":49,"value":704}," CT_BODY",{"type":44,"tag":468,"props":706,"children":707},{"style":518},[708],{"type":49,"value":709}," --output-dir",{"type":44,"tag":468,"props":711,"children":712},{"style":518},[713],{"type":49,"value":714}," OUT_DIR\n",{"type":44,"tag":417,"props":716,"children":717},{},[718,720,726,728,734,736,742,744,750,752,758,759,765,766,772,774,780,781,787],{"type":49,"value":719},"Do not invent ",{"type":44,"tag":78,"props":721,"children":723},{"className":722},[],[724],{"type":49,"value":725},"python -m nv_segment_ctmr",{"type":49,"value":727},", ",{"type":44,"tag":78,"props":729,"children":731},{"className":730},[],[732],{"type":49,"value":733},"infer.py",{"type":49,"value":735},", or ",{"type":44,"tag":78,"props":737,"children":739},{"className":738},[],[740],{"type":49,"value":741},"Medical AI Skills run",{"type":49,"value":743}," commands. ",{"type":44,"tag":78,"props":745,"children":747},{"className":746},[],[748],{"type":49,"value":749},"PATH_TO_IMAGE.nii.gz",{"type":49,"value":751}," must be the user's supplied input path.\nFor benchmark\u002Fuser run answers, the bash block is invalid if it includes\n",{"type":44,"tag":78,"props":753,"children":755},{"className":754},[],[756],{"type":49,"value":757},"mkdir -p .workbench_data\u002Fupstreams",{"type":49,"value":727},{"type":44,"tag":78,"props":760,"children":762},{"className":761},[],[763],{"type":49,"value":764},"git clone",{"type":49,"value":727},{"type":44,"tag":78,"props":767,"children":769},{"className":768},[],[770],{"type":49,"value":771},"mkdir -p \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\"",{"type":49,"value":773},",\n",{"type":44,"tag":78,"props":775,"children":777},{"className":776},[],[778],{"type":49,"value":779},"hf download",{"type":49,"value":727},{"type":44,"tag":78,"props":782,"children":784},{"className":783},[],[785],{"type":49,"value":786},"mv \"$NV_SEGMENT_CTMR_ROOT\u002F...",{"type":49,"value":788},", or any other command that\ncreates, downloads into, or moves files inside the shared upstream checkout.",{"type":44,"tag":52,"props":790,"children":792},{"id":791},"preconditions",[793],{"type":49,"value":794},"Preconditions",{"type":44,"tag":417,"props":796,"children":797},{},[798],{"type":49,"value":799},"One-time maintainer setup only; do not include these commands in user answers\nor benchmark commands. The benchmark environment already provides the\nrepo-local upstream cache and model files.",{"type":44,"tag":417,"props":801,"children":802},{},[803,805,810],{"type":49,"value":804},"If ",{"type":44,"tag":78,"props":806,"children":808},{"className":807},[],[809],{"type":49,"value":500},{"type":49,"value":811}," already names a local bundle checkout, the wrapper\nuses it and records its current commit in the result. Otherwise, clone the\nrecommended pinned default once:",{"type":44,"tag":457,"props":813,"children":815},{"className":459,"code":814,"language":461,"meta":462,"style":462},"if [ -z \"${NV_SEGMENT_CTMR_ROOT:-}\" ]; then\n  export NV_SEGMENT_CTMR_COMMIT=f9f5f51b589e5dc9c23c453cf5138398e4084056\n  export NV_SEGMENT_CTMR_CHECKOUT=\"$HOME\u002F.cache\u002Fnvidia-skills\u002Fupstreams\u002FNV-Segment-CTMR-f9f5f51\"\n  if [ ! -d \"$NV_SEGMENT_CTMR_CHECKOUT\u002F.git\" ]; then\n    git clone https:\u002F\u002Fgithub.com\u002FNVIDIA-Medtech\u002FNV-Segment-CTMR.git \"$NV_SEGMENT_CTMR_CHECKOUT\"\n    git -C \"$NV_SEGMENT_CTMR_CHECKOUT\" checkout --detach \"$NV_SEGMENT_CTMR_COMMIT\"\n  fi\n  export NV_SEGMENT_CTMR_ROOT=\"$NV_SEGMENT_CTMR_CHECKOUT\u002FNV-Segment-CTMR\"\nfi\npython -m pip install \"monai>=1.5,\u003C1.6\" \"numpy\u003C2\" nibabel scipy typer PyYAML fire huggingface_hub pytorch-ignite einops && \\\npython -c \"import monai, nibabel, numpy\"\n\nmkdir -p \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\"\ntest -e \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\u002Fmodel.pt\" || \\\n  hf download nvidia\u002FNV-Segment-CTMR \\\n    --revision 4fb8b4a6b2532be9f1c449a3726fe5440ab4213a \\\n    --local-dir \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\u002F\"\ntest -e \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\u002Fmodel.pt\" || \\\n  mv \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\u002Fvista3d_pretrained_model\u002Fmodel.pt\" \\\n    \"$NV_SEGMENT_CTMR_ROOT\u002Fmodels\u002Fmodel.pt\"\n",[816],{"type":44,"tag":78,"props":817,"children":818},{"__ignoreMap":462},[819,862,884,918,967,998,1046,1055,1088,1097,1181,1207,1217,1249,1290,1313,1331,1357,1393,1423],{"type":44,"tag":468,"props":820,"children":821},{"class":470,"line":471},[822,828,833,838,843,847,852,857],{"type":44,"tag":468,"props":823,"children":825},{"style":824},"--shiki-light:#39ADB5;--shiki-light-font-style:italic;--shiki-default:#89DDFF;--shiki-default-font-style:italic;--shiki-dark:#89DDFF;--shiki-dark-font-style:italic",[826],{"type":49,"value":827},"if",{"type":44,"tag":468,"props":829,"children":830},{"style":487},[831],{"type":49,"value":832}," [",{"type":44,"tag":468,"props":834,"children":835},{"style":487},[836],{"type":49,"value":837}," -z",{"type":44,"tag":468,"props":839,"children":840},{"style":487},[841],{"type":49,"value":842}," \"${",{"type":44,"tag":468,"props":844,"children":845},{"style":481},[846],{"type":49,"value":500},{"type":44,"tag":468,"props":848,"children":849},{"style":487},[850],{"type":49,"value":851},":-}\"",{"type":44,"tag":468,"props":853,"children":854},{"style":487},[855],{"type":49,"value":856}," ];",{"type":44,"tag":468,"props":858,"children":859},{"style":824},[860],{"type":49,"value":861}," then\n",{"type":44,"tag":468,"props":863,"children":864},{"class":470,"line":578},[865,870,875,879],{"type":44,"tag":468,"props":866,"children":867},{"style":475},[868],{"type":49,"value":869},"  export",{"type":44,"tag":468,"props":871,"children":872},{"style":481},[873],{"type":49,"value":874}," NV_SEGMENT_CTMR_COMMIT",{"type":44,"tag":468,"props":876,"children":877},{"style":487},[878],{"type":49,"value":490},{"type":44,"tag":468,"props":880,"children":881},{"style":481},[882],{"type":49,"value":883},"f9f5f51b589e5dc9c23c453cf5138398e4084056\n",{"type":44,"tag":468,"props":885,"children":886},{"class":470,"line":679},[887,891,896,900,904,908,913],{"type":44,"tag":468,"props":888,"children":889},{"style":475},[890],{"type":49,"value":869},{"type":44,"tag":468,"props":892,"children":893},{"style":481},[894],{"type":49,"value":895}," NV_SEGMENT_CTMR_CHECKOUT",{"type":44,"tag":468,"props":897,"children":898},{"style":487},[899],{"type":49,"value":490},{"type":44,"tag":468,"props":901,"children":902},{"style":487},[903],{"type":49,"value":615},{"type":44,"tag":468,"props":905,"children":906},{"style":481},[907],{"type":49,"value":510},{"type":44,"tag":468,"props":909,"children":910},{"style":518},[911],{"type":49,"value":912},"\u002F.cache\u002Fnvidia-skills\u002Fupstreams\u002FNV-Segment-CTMR-f9f5f51",{"type":44,"tag":468,"props":914,"children":915},{"style":487},[916],{"type":49,"value":917},"\"\n",{"type":44,"tag":468,"props":919,"children":921},{"class":470,"line":920},4,[922,927,931,936,941,945,950,955,959,963],{"type":44,"tag":468,"props":923,"children":924},{"style":824},[925],{"type":49,"value":926},"  if",{"type":44,"tag":468,"props":928,"children":929},{"style":487},[930],{"type":49,"value":832},{"type":44,"tag":468,"props":932,"children":933},{"style":487},[934],{"type":49,"value":935}," !",{"type":44,"tag":468,"props":937,"children":938},{"style":487},[939],{"type":49,"value":940}," -d",{"type":44,"tag":468,"props":942,"children":943},{"style":487},[944],{"type":49,"value":605},{"type":44,"tag":468,"props":946,"children":947},{"style":481},[948],{"type":49,"value":949},"$NV_SEGMENT_CTMR_CHECKOUT",{"type":44,"tag":468,"props":951,"children":952},{"style":518},[953],{"type":49,"value":954},"\u002F.git",{"type":44,"tag":468,"props":956,"children":957},{"style":487},[958],{"type":49,"value":615},{"type":44,"tag":468,"props":960,"children":961},{"style":487},[962],{"type":49,"value":856},{"type":44,"tag":468,"props":964,"children":965},{"style":824},[966],{"type":49,"value":861},{"type":44,"tag":468,"props":968,"children":970},{"class":470,"line":969},5,[971,976,981,986,990,994],{"type":44,"tag":468,"props":972,"children":973},{"style":582},[974],{"type":49,"value":975},"    git",{"type":44,"tag":468,"props":977,"children":978},{"style":518},[979],{"type":49,"value":980}," clone",{"type":44,"tag":468,"props":982,"children":983},{"style":518},[984],{"type":49,"value":985}," https:\u002F\u002Fgithub.com\u002FNVIDIA-Medtech\u002FNV-Segment-CTMR.git",{"type":44,"tag":468,"props":987,"children":988},{"style":487},[989],{"type":49,"value":605},{"type":44,"tag":468,"props":991,"children":992},{"style":481},[993],{"type":49,"value":949},{"type":44,"tag":468,"props":995,"children":996},{"style":487},[997],{"type":49,"value":917},{"type":44,"tag":468,"props":999,"children":1001},{"class":470,"line":1000},6,[1002,1006,1011,1015,1019,1023,1028,1033,1037,1042],{"type":44,"tag":468,"props":1003,"children":1004},{"style":582},[1005],{"type":49,"value":975},{"type":44,"tag":468,"props":1007,"children":1008},{"style":518},[1009],{"type":49,"value":1010}," -C",{"type":44,"tag":468,"props":1012,"children":1013},{"style":487},[1014],{"type":49,"value":605},{"type":44,"tag":468,"props":1016,"children":1017},{"style":481},[1018],{"type":49,"value":949},{"type":44,"tag":468,"props":1020,"children":1021},{"style":487},[1022],{"type":49,"value":615},{"type":44,"tag":468,"props":1024,"children":1025},{"style":518},[1026],{"type":49,"value":1027}," checkout",{"type":44,"tag":468,"props":1029,"children":1030},{"style":518},[1031],{"type":49,"value":1032}," --detach",{"type":44,"tag":468,"props":1034,"children":1035},{"style":487},[1036],{"type":49,"value":605},{"type":44,"tag":468,"props":1038,"children":1039},{"style":481},[1040],{"type":49,"value":1041},"$NV_SEGMENT_CTMR_COMMIT",{"type":44,"tag":468,"props":1043,"children":1044},{"style":487},[1045],{"type":49,"value":917},{"type":44,"tag":468,"props":1047,"children":1049},{"class":470,"line":1048},7,[1050],{"type":44,"tag":468,"props":1051,"children":1052},{"style":824},[1053],{"type":49,"value":1054},"  fi\n",{"type":44,"tag":468,"props":1056,"children":1058},{"class":470,"line":1057},8,[1059,1063,1067,1071,1075,1079,1084],{"type":44,"tag":468,"props":1060,"children":1061},{"style":475},[1062],{"type":49,"value":869},{"type":44,"tag":468,"props":1064,"children":1065},{"style":481},[1066],{"type":49,"value":484},{"type":44,"tag":468,"props":1068,"children":1069},{"style":487},[1070],{"type":49,"value":490},{"type":44,"tag":468,"props":1072,"children":1073},{"style":487},[1074],{"type":49,"value":615},{"type":44,"tag":468,"props":1076,"children":1077},{"style":481},[1078],{"type":49,"value":949},{"type":44,"tag":468,"props":1080,"children":1081},{"style":518},[1082],{"type":49,"value":1083},"\u002FNV-Segment-CTMR",{"type":44,"tag":468,"props":1085,"children":1086},{"style":487},[1087],{"type":49,"value":917},{"type":44,"tag":468,"props":1089,"children":1091},{"class":470,"line":1090},9,[1092],{"type":44,"tag":468,"props":1093,"children":1094},{"style":824},[1095],{"type":49,"value":1096},"fi\n",{"type":44,"tag":468,"props":1098,"children":1100},{"class":470,"line":1099},10,[1101,1105,1109,1113,1117,1121,1125,1129,1133,1137,1141,1145,1149,1153,1157,1161,1165,1169,1173,1177],{"type":44,"tag":468,"props":1102,"children":1103},{"style":582},[1104],{"type":49,"value":585},{"type":44,"tag":468,"props":1106,"children":1107},{"style":518},[1108],{"type":49,"value":590},{"type":44,"tag":468,"props":1110,"children":1111},{"style":518},[1112],{"type":49,"value":595},{"type":44,"tag":468,"props":1114,"children":1115},{"style":518},[1116],{"type":49,"value":600},{"type":44,"tag":468,"props":1118,"children":1119},{"style":487},[1120],{"type":49,"value":605},{"type":44,"tag":468,"props":1122,"children":1123},{"style":518},[1124],{"type":49,"value":610},{"type":44,"tag":468,"props":1126,"children":1127},{"style":487},[1128],{"type":49,"value":615},{"type":44,"tag":468,"props":1130,"children":1131},{"style":487},[1132],{"type":49,"value":605},{"type":44,"tag":468,"props":1134,"children":1135},{"style":518},[1136],{"type":49,"value":624},{"type":44,"tag":468,"props":1138,"children":1139},{"style":487},[1140],{"type":49,"value":615},{"type":44,"tag":468,"props":1142,"children":1143},{"style":518},[1144],{"type":49,"value":633},{"type":44,"tag":468,"props":1146,"children":1147},{"style":518},[1148],{"type":49,"value":638},{"type":44,"tag":468,"props":1150,"children":1151},{"style":518},[1152],{"type":49,"value":643},{"type":44,"tag":468,"props":1154,"children":1155},{"style":518},[1156],{"type":49,"value":648},{"type":44,"tag":468,"props":1158,"children":1159},{"style":518},[1160],{"type":49,"value":653},{"type":44,"tag":468,"props":1162,"children":1163},{"style":518},[1164],{"type":49,"value":658},{"type":44,"tag":468,"props":1166,"children":1167},{"style":518},[1168],{"type":49,"value":663},{"type":44,"tag":468,"props":1170,"children":1171},{"style":518},[1172],{"type":49,"value":668},{"type":44,"tag":468,"props":1174,"children":1175},{"style":487},[1176],{"type":49,"value":570},{"type":44,"tag":468,"props":1178,"children":1179},{"style":481},[1180],{"type":49,"value":575},{"type":44,"tag":468,"props":1182,"children":1184},{"class":470,"line":1183},11,[1185,1189,1194,1198,1203],{"type":44,"tag":468,"props":1186,"children":1187},{"style":582},[1188],{"type":49,"value":585},{"type":44,"tag":468,"props":1190,"children":1191},{"style":518},[1192],{"type":49,"value":1193}," -c",{"type":44,"tag":468,"props":1195,"children":1196},{"style":487},[1197],{"type":49,"value":605},{"type":44,"tag":468,"props":1199,"children":1200},{"style":518},[1201],{"type":49,"value":1202},"import monai, nibabel, numpy",{"type":44,"tag":468,"props":1204,"children":1205},{"style":487},[1206],{"type":49,"value":917},{"type":44,"tag":468,"props":1208,"children":1210},{"class":470,"line":1209},12,[1211],{"type":44,"tag":468,"props":1212,"children":1214},{"emptyLinePlaceholder":1213},true,[1215],{"type":49,"value":1216},"\n",{"type":44,"tag":468,"props":1218,"children":1220},{"class":470,"line":1219},13,[1221,1226,1231,1235,1240,1245],{"type":44,"tag":468,"props":1222,"children":1223},{"style":582},[1224],{"type":49,"value":1225},"mkdir",{"type":44,"tag":468,"props":1227,"children":1228},{"style":518},[1229],{"type":49,"value":1230}," -p",{"type":44,"tag":468,"props":1232,"children":1233},{"style":487},[1234],{"type":49,"value":605},{"type":44,"tag":468,"props":1236,"children":1237},{"style":481},[1238],{"type":49,"value":1239},"$NV_SEGMENT_CTMR_ROOT",{"type":44,"tag":468,"props":1241,"children":1242},{"style":518},[1243],{"type":49,"value":1244},"\u002Fmodels",{"type":44,"tag":468,"props":1246,"children":1247},{"style":487},[1248],{"type":49,"value":917},{"type":44,"tag":468,"props":1250,"children":1252},{"class":470,"line":1251},14,[1253,1259,1264,1268,1272,1277,1281,1286],{"type":44,"tag":468,"props":1254,"children":1256},{"style":1255},"--shiki-light:#6182B8;--shiki-default:#82AAFF;--shiki-dark:#82AAFF",[1257],{"type":49,"value":1258},"test",{"type":44,"tag":468,"props":1260,"children":1261},{"style":518},[1262],{"type":49,"value":1263}," -e",{"type":44,"tag":468,"props":1265,"children":1266},{"style":487},[1267],{"type":49,"value":605},{"type":44,"tag":468,"props":1269,"children":1270},{"style":481},[1271],{"type":49,"value":1239},{"type":44,"tag":468,"props":1273,"children":1274},{"style":518},[1275],{"type":49,"value":1276},"\u002Fmodels\u002Fmodel.pt",{"type":44,"tag":468,"props":1278,"children":1279},{"style":487},[1280],{"type":49,"value":615},{"type":44,"tag":468,"props":1282,"children":1283},{"style":487},[1284],{"type":49,"value":1285}," ||",{"type":44,"tag":468,"props":1287,"children":1288},{"style":481},[1289],{"type":49,"value":575},{"type":44,"tag":468,"props":1291,"children":1293},{"class":470,"line":1292},15,[1294,1299,1304,1309],{"type":44,"tag":468,"props":1295,"children":1296},{"style":582},[1297],{"type":49,"value":1298},"  hf",{"type":44,"tag":468,"props":1300,"children":1301},{"style":518},[1302],{"type":49,"value":1303}," download",{"type":44,"tag":468,"props":1305,"children":1306},{"style":518},[1307],{"type":49,"value":1308}," nvidia\u002FNV-Segment-CTMR",{"type":44,"tag":468,"props":1310,"children":1311},{"style":481},[1312],{"type":49,"value":575},{"type":44,"tag":468,"props":1314,"children":1316},{"class":470,"line":1315},16,[1317,1322,1327],{"type":44,"tag":468,"props":1318,"children":1319},{"style":518},[1320],{"type":49,"value":1321},"    --revision",{"type":44,"tag":468,"props":1323,"children":1324},{"style":518},[1325],{"type":49,"value":1326}," 4fb8b4a6b2532be9f1c449a3726fe5440ab4213a",{"type":44,"tag":468,"props":1328,"children":1329},{"style":481},[1330],{"type":49,"value":575},{"type":44,"tag":468,"props":1332,"children":1334},{"class":470,"line":1333},17,[1335,1340,1344,1348,1353],{"type":44,"tag":468,"props":1336,"children":1337},{"style":518},[1338],{"type":49,"value":1339},"    --local-dir",{"type":44,"tag":468,"props":1341,"children":1342},{"style":487},[1343],{"type":49,"value":605},{"type":44,"tag":468,"props":1345,"children":1346},{"style":481},[1347],{"type":49,"value":1239},{"type":44,"tag":468,"props":1349,"children":1350},{"style":518},[1351],{"type":49,"value":1352},"\u002Fmodels\u002F",{"type":44,"tag":468,"props":1354,"children":1355},{"style":487},[1356],{"type":49,"value":917},{"type":44,"tag":468,"props":1358,"children":1360},{"class":470,"line":1359},18,[1361,1365,1369,1373,1377,1381,1385,1389],{"type":44,"tag":468,"props":1362,"children":1363},{"style":1255},[1364],{"type":49,"value":1258},{"type":44,"tag":468,"props":1366,"children":1367},{"style":518},[1368],{"type":49,"value":1263},{"type":44,"tag":468,"props":1370,"children":1371},{"style":487},[1372],{"type":49,"value":605},{"type":44,"tag":468,"props":1374,"children":1375},{"style":481},[1376],{"type":49,"value":1239},{"type":44,"tag":468,"props":1378,"children":1379},{"style":518},[1380],{"type":49,"value":1276},{"type":44,"tag":468,"props":1382,"children":1383},{"style":487},[1384],{"type":49,"value":615},{"type":44,"tag":468,"props":1386,"children":1387},{"style":487},[1388],{"type":49,"value":1285},{"type":44,"tag":468,"props":1390,"children":1391},{"style":481},[1392],{"type":49,"value":575},{"type":44,"tag":468,"props":1394,"children":1396},{"class":470,"line":1395},19,[1397,1402,1406,1410,1415,1419],{"type":44,"tag":468,"props":1398,"children":1399},{"style":582},[1400],{"type":49,"value":1401},"  mv",{"type":44,"tag":468,"props":1403,"children":1404},{"style":487},[1405],{"type":49,"value":605},{"type":44,"tag":468,"props":1407,"children":1408},{"style":481},[1409],{"type":49,"value":1239},{"type":44,"tag":468,"props":1411,"children":1412},{"style":518},[1413],{"type":49,"value":1414},"\u002Fmodels\u002Fvista3d_pretrained_model\u002Fmodel.pt",{"type":44,"tag":468,"props":1416,"children":1417},{"style":487},[1418],{"type":49,"value":615},{"type":44,"tag":468,"props":1420,"children":1421},{"style":481},[1422],{"type":49,"value":575},{"type":44,"tag":468,"props":1424,"children":1426},{"class":470,"line":1425},20,[1427,1432,1436,1440],{"type":44,"tag":468,"props":1428,"children":1429},{"style":487},[1430],{"type":49,"value":1431},"    \"",{"type":44,"tag":468,"props":1433,"children":1434},{"style":481},[1435],{"type":49,"value":1239},{"type":44,"tag":468,"props":1437,"children":1438},{"style":518},[1439],{"type":49,"value":1276},{"type":44,"tag":468,"props":1441,"children":1442},{"style":487},[1443],{"type":49,"value":917},{"type":44,"tag":417,"props":1445,"children":1446},{},[1447,1449,1455,1457,1462],{"type":49,"value":1448},"The wrapper also searches ",{"type":44,"tag":78,"props":1450,"children":1452},{"className":1451},[],[1453],{"type":49,"value":1454},".workbench_data\u002Fupstreams\u002FNV-Segment-CTMR\u002FNV-Segment-CTMR",{"type":49,"value":1456},"\nif ",{"type":44,"tag":78,"props":1458,"children":1460},{"className":1459},[],[1461],{"type":49,"value":500},{"type":49,"value":1463}," is unset or does not have the required bundle layout.",{"type":44,"tag":417,"props":1465,"children":1466},{},[1467,1469,1474,1476,1481,1483,1489,1491,1496,1497,1502,1503,1509,1510,1516,1517,1523,1524,1530,1531,1537,1538,1544,1545,1551,1552,1558],{"type":49,"value":1468},"For agent-generated user run commands, use the command in Usage. Do not copy\nthe one-time Preconditions block into the answer: do not create or write under\n",{"type":44,"tag":78,"props":1470,"children":1472},{"className":1471},[],[1473],{"type":49,"value":1239},{"type":49,"value":1475},", do not run ",{"type":44,"tag":78,"props":1477,"children":1479},{"className":1478},[],[1480],{"type":49,"value":779},{"type":49,"value":1482},", and do not move files in the\nshared upstream checkout during a benchmark or user run. Do not prepend\n",{"type":44,"tag":78,"props":1484,"children":1486},{"className":1485},[],[1487],{"type":49,"value":1488},"pip install -r \"$NV_SEGMENT_CTMR_ROOT\u002Frequirements.txt\"",{"type":49,"value":1490}," in a Python 3.12\nenvironment; the upstream requirements pin NumPy 1.24.4, which does not build\ncleanly there. In a fresh Python environment, install the minimal compatible\nruntime shown above (",{"type":44,"tag":78,"props":1492,"children":1494},{"className":1493},[],[1495],{"type":49,"value":610},{"type":49,"value":727},{"type":44,"tag":78,"props":1498,"children":1500},{"className":1499},[],[1501],{"type":49,"value":624},{"type":49,"value":727},{"type":44,"tag":78,"props":1504,"children":1506},{"className":1505},[],[1507],{"type":49,"value":1508},"nibabel",{"type":49,"value":727},{"type":44,"tag":78,"props":1511,"children":1513},{"className":1512},[],[1514],{"type":49,"value":1515},"scipy",{"type":49,"value":727},{"type":44,"tag":78,"props":1518,"children":1520},{"className":1519},[],[1521],{"type":49,"value":1522},"typer",{"type":49,"value":773},{"type":44,"tag":78,"props":1525,"children":1527},{"className":1526},[],[1528],{"type":49,"value":1529},"PyYAML",{"type":49,"value":727},{"type":44,"tag":78,"props":1532,"children":1534},{"className":1533},[],[1535],{"type":49,"value":1536},"fire",{"type":49,"value":727},{"type":44,"tag":78,"props":1539,"children":1541},{"className":1540},[],[1542],{"type":49,"value":1543},"huggingface_hub",{"type":49,"value":727},{"type":44,"tag":78,"props":1546,"children":1548},{"className":1547},[],[1549],{"type":49,"value":1550},"pytorch-ignite",{"type":49,"value":727},{"type":44,"tag":78,"props":1553,"children":1555},{"className":1554},[],[1556],{"type":49,"value":1557},"einops",{"type":49,"value":1559},") before the\nwrapper. Cached models do not imply cached Python packages.",{"type":44,"tag":417,"props":1561,"children":1562},{},[1563],{"type":49,"value":1564},"Runtime needs an NVIDIA GPU with CUDA. The upstream bundle may import on\nCPU-only hosts, but this skill is declared as CUDA-required because the\npublished workflow is a 3D CT\u002FMRI foundation model inference path.",{"type":44,"tag":52,"props":1566,"children":1568},{"id":1567},"usage",[1569],{"type":49,"value":1570},"Usage",{"type":44,"tag":417,"props":1572,"children":1573},{},[1574],{"type":49,"value":1575},"From Medical AI Skills repo root:",{"type":44,"tag":457,"props":1577,"children":1579},{"className":459,"code":1578,"language":461,"meta":462,"style":462},"export NV_SEGMENT_CTMR_ROOT=\"${NV_SEGMENT_CTMR_ROOT:-$HOME\u002F.cache\u002Fnvidia-skills\u002Fupstreams\u002FNV-Segment-CTMR-f9f5f51\u002FNV-Segment-CTMR}\" && \\\npython -m pip install \"monai>=1.5,\u003C1.6\" \"numpy\u003C2\" nibabel scipy typer PyYAML fire huggingface_hub pytorch-ignite einops && \\\npython skills\u002Fnv-segment-ctmr\u002Fscripts\u002Frun_ctmr.py PATH_TO_IMAGE.nii.gz \\\n  --modality CT_BODY \\\n  --output-dir runs\u002Fnv_segment_ctmr_demo\n",[1580],{"type":44,"tag":78,"props":1581,"children":1582},{"__ignoreMap":462},[1583,1670,1753,1772,1788],{"type":44,"tag":468,"props":1584,"children":1585},{"class":470,"line":471},[1586,1590,1594,1598,1602,1606,1610,1614,1618,1622,1626,1630,1634,1638,1642,1646,1650,1654,1658,1662,1666],{"type":44,"tag":468,"props":1587,"children":1588},{"style":475},[1589],{"type":49,"value":478},{"type":44,"tag":468,"props":1591,"children":1592},{"style":481},[1593],{"type":49,"value":484},{"type":44,"tag":468,"props":1595,"children":1596},{"style":487},[1597],{"type":49,"value":490},{"type":44,"tag":468,"props":1599,"children":1600},{"style":487},[1601],{"type":49,"value":495},{"type":44,"tag":468,"props":1603,"children":1604},{"style":481},[1605],{"type":49,"value":500},{"type":44,"tag":468,"props":1607,"children":1608},{"style":487},[1609],{"type":49,"value":505},{"type":44,"tag":468,"props":1611,"children":1612},{"style":481},[1613],{"type":49,"value":510},{"type":44,"tag":468,"props":1615,"children":1616},{"style":487},[1617],{"type":49,"value":515},{"type":44,"tag":468,"props":1619,"children":1620},{"style":518},[1621],{"type":49,"value":101},{"type":44,"tag":468,"props":1623,"children":1624},{"style":481},[1625],{"type":49,"value":525},{"type":44,"tag":468,"props":1627,"children":1628},{"style":487},[1629],{"type":49,"value":515},{"type":44,"tag":468,"props":1631,"children":1632},{"style":481},[1633],{"type":49,"value":534},{"type":44,"tag":468,"props":1635,"children":1636},{"style":487},[1637],{"type":49,"value":515},{"type":44,"tag":468,"props":1639,"children":1640},{"style":481},[1641],{"type":49,"value":543},{"type":44,"tag":468,"props":1643,"children":1644},{"style":487},[1645],{"type":49,"value":515},{"type":44,"tag":468,"props":1647,"children":1648},{"style":481},[1649],{"type":49,"value":552},{"type":44,"tag":468,"props":1651,"children":1652},{"style":487},[1653],{"type":49,"value":515},{"type":44,"tag":468,"props":1655,"children":1656},{"style":481},[1657],{"type":49,"value":50},{"type":44,"tag":468,"props":1659,"children":1660},{"style":487},[1661],{"type":49,"value":565},{"type":44,"tag":468,"props":1663,"children":1664},{"style":487},[1665],{"type":49,"value":570},{"type":44,"tag":468,"props":1667,"children":1668},{"style":481},[1669],{"type":49,"value":575},{"type":44,"tag":468,"props":1671,"children":1672},{"class":470,"line":578},[1673,1677,1681,1685,1689,1693,1697,1701,1705,1709,1713,1717,1721,1725,1729,1733,1737,1741,1745,1749],{"type":44,"tag":468,"props":1674,"children":1675},{"style":582},[1676],{"type":49,"value":585},{"type":44,"tag":468,"props":1678,"children":1679},{"style":518},[1680],{"type":49,"value":590},{"type":44,"tag":468,"props":1682,"children":1683},{"style":518},[1684],{"type":49,"value":595},{"type":44,"tag":468,"props":1686,"children":1687},{"style":518},[1688],{"type":49,"value":600},{"type":44,"tag":468,"props":1690,"children":1691},{"style":487},[1692],{"type":49,"value":605},{"type":44,"tag":468,"props":1694,"children":1695},{"style":518},[1696],{"type":49,"value":610},{"type":44,"tag":468,"props":1698,"children":1699},{"style":487},[1700],{"type":49,"value":615},{"type":44,"tag":468,"props":1702,"children":1703},{"style":487},[1704],{"type":49,"value":605},{"type":44,"tag":468,"props":1706,"children":1707},{"style":518},[1708],{"type":49,"value":624},{"type":44,"tag":468,"props":1710,"children":1711},{"style":487},[1712],{"type":49,"value":615},{"type":44,"tag":468,"props":1714,"children":1715},{"style":518},[1716],{"type":49,"value":633},{"type":44,"tag":468,"props":1718,"children":1719},{"style":518},[1720],{"type":49,"value":638},{"type":44,"tag":468,"props":1722,"children":1723},{"style":518},[1724],{"type":49,"value":643},{"type":44,"tag":468,"props":1726,"children":1727},{"style":518},[1728],{"type":49,"value":648},{"type":44,"tag":468,"props":1730,"children":1731},{"style":518},[1732],{"type":49,"value":653},{"type":44,"tag":468,"props":1734,"children":1735},{"style":518},[1736],{"type":49,"value":658},{"type":44,"tag":468,"props":1738,"children":1739},{"style":518},[1740],{"type":49,"value":663},{"type":44,"tag":468,"props":1742,"children":1743},{"style":518},[1744],{"type":49,"value":668},{"type":44,"tag":468,"props":1746,"children":1747},{"style":487},[1748],{"type":49,"value":570},{"type":44,"tag":468,"props":1750,"children":1751},{"style":481},[1752],{"type":49,"value":575},{"type":44,"tag":468,"props":1754,"children":1755},{"class":470,"line":679},[1756,1760,1764,1768],{"type":44,"tag":468,"props":1757,"children":1758},{"style":582},[1759],{"type":49,"value":585},{"type":44,"tag":468,"props":1761,"children":1762},{"style":518},[1763],{"type":49,"value":689},{"type":44,"tag":468,"props":1765,"children":1766},{"style":518},[1767],{"type":49,"value":694},{"type":44,"tag":468,"props":1769,"children":1770},{"style":481},[1771],{"type":49,"value":575},{"type":44,"tag":468,"props":1773,"children":1774},{"class":470,"line":920},[1775,1780,1784],{"type":44,"tag":468,"props":1776,"children":1777},{"style":518},[1778],{"type":49,"value":1779},"  --modality",{"type":44,"tag":468,"props":1781,"children":1782},{"style":518},[1783],{"type":49,"value":704},{"type":44,"tag":468,"props":1785,"children":1786},{"style":481},[1787],{"type":49,"value":575},{"type":44,"tag":468,"props":1789,"children":1790},{"class":470,"line":969},[1791,1796],{"type":44,"tag":468,"props":1792,"children":1793},{"style":518},[1794],{"type":49,"value":1795},"  --output-dir",{"type":44,"tag":468,"props":1797,"children":1798},{"style":518},[1799],{"type":49,"value":1800}," runs\u002Fnv_segment_ctmr_demo\n",{"type":44,"tag":417,"props":1802,"children":1803},{},[1804,1806,1811,1813,1819,1821,1826],{"type":49,"value":1805},"Replace ",{"type":44,"tag":78,"props":1807,"children":1809},{"className":1808},[],[1810],{"type":49,"value":749},{"type":49,"value":1812}," with the user's actual input path. Do not copy\nthe example fixture path into a user run. If the user provides an explicit\ninput path under ",{"type":44,"tag":78,"props":1814,"children":1816},{"className":1815},[],[1817],{"type":49,"value":1818},"runs\u002F",{"type":49,"value":1820},", that path must be the first positional argument to\n",{"type":44,"tag":78,"props":1822,"children":1824},{"className":1823},[],[1825],{"type":49,"value":134},{"type":49,"value":101},{"type":44,"tag":417,"props":1828,"children":1829},{},[1830,1832,1838,1839,1845,1847,1853,1855,1860],{"type":49,"value":1831},"Supported automatic segmentation modalities are ",{"type":44,"tag":78,"props":1833,"children":1835},{"className":1834},[],[1836],{"type":49,"value":1837},"CT_BODY",{"type":49,"value":727},{"type":44,"tag":78,"props":1840,"children":1842},{"className":1841},[],[1843],{"type":49,"value":1844},"MRI_BODY",{"type":49,"value":1846},", and\n",{"type":44,"tag":78,"props":1848,"children":1850},{"className":1849},[],[1851],{"type":49,"value":1852},"MRI_BRAIN",{"type":49,"value":1854},". For ",{"type":44,"tag":78,"props":1856,"children":1858},{"className":1857},[],[1859],{"type":49,"value":1852},{"type":49,"value":1861},", the upstream README requires brain-specific\npreprocessing before bundle inference; pass an already preprocessed image to\nthis wrapper.",{"type":44,"tag":417,"props":1863,"children":1864},{},[1865,1867,1873],{"type":49,"value":1866},"Pass ",{"type":44,"tag":78,"props":1868,"children":1870},{"className":1869},[],[1871],{"type":49,"value":1872},"--label-prompts \"3,14\"",{"type":49,"value":1874}," to request specific upstream class IDs instead\nof only the modality-level \"segment everything\" set. The evidence output\nrecords input geometry, output mask path, observed label IDs, unexpected\nlabels, per-class voxel counts, per-class physical volumes from the mask\nheader spacing, runtime, upstream command, model inventory, and geometry\nchecks.",{"type":44,"tag":417,"props":1876,"children":1877},{},[1878,1879,1885,1887,1893],{"type":49,"value":1866},{"type":44,"tag":78,"props":1880,"children":1882},{"className":1881},[],[1883],{"type":49,"value":1884},"--ground-truth PATH",{"type":49,"value":1886}," to record a reference label-map path under\n",{"type":44,"tag":78,"props":1888,"children":1890},{"className":1889},[],[1891],{"type":49,"value":1892},"input.ground_truth_path",{"type":49,"value":1894},". The skill does not compute Dice; that is the\npaired verifier's job.",{"type":44,"tag":417,"props":1896,"children":1897},{},[1898,1900,1906],{"type":49,"value":1899},"Anatomy plausibility and optional per-class Dice\u002FIoU against the recorded\nground truth can be checked by ",{"type":44,"tag":78,"props":1901,"children":1903},{"className":1902},[],[1904],{"type":49,"value":1905},"verifiers\u002Fct_segmentation_quality_v1",{"type":49,"value":1907}," for\nCT-body outputs.",{"type":44,"tag":417,"props":1909,"children":1910},{},[1911],{"type":49,"value":1912},"Not for clinical interpretation, production deployment, autonomous diagnosis,\nor regulatory submission.",{"type":44,"tag":1914,"props":1915,"children":1916},"style",{},[1917],{"type":49,"value":1918},"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":1920,"total":2078},[1921,1939,1956,1967,1979,1993,2006,2020,2033,2044,2058,2067],{"slug":1922,"name":1922,"fn":1923,"description":1924,"org":1925,"tags":1926,"stars":1936,"repoUrl":1937,"updatedAt":1938},"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},[1927,1930,1933],{"name":1928,"slug":1929,"type":15},"Documentation","documentation",{"name":1931,"slug":1932,"type":15},"MCP","mcp",{"name":1934,"slug":1935,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":1940,"name":1940,"fn":1941,"description":1942,"org":1943,"tags":1944,"stars":1953,"repoUrl":1954,"updatedAt":1955},"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},[1945,1948,1951],{"name":1946,"slug":1947,"type":15},"Containers","containers",{"name":1949,"slug":1950,"type":15},"Deployment","deployment",{"name":1952,"slug":585,"type":15},"Python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":1957,"name":1957,"fn":1958,"description":1959,"org":1960,"tags":1961,"stars":1953,"repoUrl":1954,"updatedAt":1966},"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},[1962,1965],{"name":1963,"slug":1964,"type":15},"CI\u002FCD","ci-cd",{"name":1949,"slug":1950,"type":15},"2026-07-14T05:25:59.97109",{"slug":1968,"name":1968,"fn":1969,"description":1970,"org":1971,"tags":1972,"stars":1953,"repoUrl":1954,"updatedAt":1978},"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},[1973,1974,1975],{"name":1963,"slug":1964,"type":15},{"name":1949,"slug":1950,"type":15},{"name":1976,"slug":1977,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":1980,"name":1980,"fn":1981,"description":1982,"org":1983,"tags":1984,"stars":1953,"repoUrl":1954,"updatedAt":1992},"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},[1985,1988,1989],{"name":1986,"slug":1987,"type":15},"Debugging","debugging",{"name":1976,"slug":1977,"type":15},{"name":1990,"slug":1991,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":1994,"name":1994,"fn":1995,"description":1996,"org":1997,"tags":1998,"stars":1953,"repoUrl":1954,"updatedAt":2005},"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},[1999,2002],{"name":2000,"slug":2001,"type":15},"Best Practices","best-practices",{"name":2003,"slug":2004,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":2007,"name":2007,"fn":2008,"description":2009,"org":2010,"tags":2011,"stars":1953,"repoUrl":1954,"updatedAt":2019},"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},[2012,2015,2018],{"name":2013,"slug":2014,"type":15},"Machine Learning","machine-learning",{"name":2016,"slug":2017,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":2021,"name":2021,"fn":2022,"description":2023,"org":2024,"tags":2025,"stars":1953,"repoUrl":1954,"updatedAt":2032},"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},[2026,2029],{"name":2027,"slug":2028,"type":15},"QA","qa",{"name":2030,"slug":2031,"type":15},"Testing","testing","2026-07-14T05:25:53.673039",{"slug":2034,"name":2034,"fn":2035,"description":2036,"org":2037,"tags":2038,"stars":1953,"repoUrl":1954,"updatedAt":2043},"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},[2039,2040],{"name":1949,"slug":1950,"type":15},{"name":2041,"slug":2042,"type":15},"Infrastructure","infrastructure","2026-07-14T05:25:49.362534",{"slug":2045,"name":2045,"fn":2046,"description":2047,"org":2048,"tags":2049,"stars":1953,"repoUrl":1954,"updatedAt":2057},"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},[2050,2053,2054],{"name":2051,"slug":2052,"type":15},"Code Review","code-review",{"name":1976,"slug":1977,"type":15},{"name":2055,"slug":2056,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":2059,"name":2059,"fn":2060,"description":2061,"org":2062,"tags":2063,"stars":1953,"repoUrl":1954,"updatedAt":2066},"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},[2064,2065],{"name":2027,"slug":2028,"type":15},{"name":2030,"slug":2031,"type":15},"2026-07-14T05:25:54.928983",{"slug":2068,"name":2068,"fn":2069,"description":2070,"org":2071,"tags":2072,"stars":1953,"repoUrl":1954,"updatedAt":2077},"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},[2073,2076],{"name":2074,"slug":2075,"type":15},"Automation","automation",{"name":1963,"slug":1964,"type":15},"2026-07-30T05:29:03.275638",496,{"items":2080,"total":2174},[2081,2098,2108,2122,2132,2145,2160],{"slug":2082,"name":2082,"fn":2083,"description":2084,"org":2085,"tags":2086,"stars":20,"repoUrl":21,"updatedAt":2097},"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},[2087,2090,2093,2094],{"name":2088,"slug":2089,"type":15},"Data Analysis","data-analysis",{"name":2091,"slug":2092,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":2095,"slug":2096,"type":15},"Performance","performance","2026-07-14T05:28:43.176466",{"slug":2099,"name":2099,"fn":2100,"description":2101,"org":2102,"tags":2103,"stars":20,"repoUrl":21,"updatedAt":2107},"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},[2104,2105,2106],{"name":1949,"slug":1950,"type":15},{"name":2041,"slug":2042,"type":15},{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":2109,"name":2109,"fn":2110,"description":2111,"org":2112,"tags":2113,"stars":20,"repoUrl":21,"updatedAt":2121},"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},[2114,2117,2118],{"name":2115,"slug":2116,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":2119,"slug":2120,"type":15},"Research","research","2026-07-14T05:28:06.816956",{"slug":2123,"name":2123,"fn":2124,"description":2125,"org":2126,"tags":2127,"stars":20,"repoUrl":21,"updatedAt":2131},"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},[2128,2129,2130],{"name":2088,"slug":2089,"type":15},{"name":9,"slug":8,"type":15},{"name":2030,"slug":2031,"type":15},"2026-07-17T05:29:03.913266",{"slug":2133,"name":2133,"fn":2134,"description":2135,"org":2136,"tags":2137,"stars":20,"repoUrl":21,"updatedAt":2144},"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},[2138,2139,2140,2141],{"name":2074,"slug":2075,"type":15},{"name":18,"slug":19,"type":15},{"name":9,"slug":8,"type":15},{"name":2142,"slug":2143,"type":15},"Video","video","2026-07-17T05:28:53.905004",{"slug":2146,"name":2146,"fn":2147,"description":2148,"org":2149,"tags":2150,"stars":20,"repoUrl":21,"updatedAt":2159},"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},[2151,2152,2155,2156],{"name":1949,"slug":1950,"type":15},{"name":2153,"slug":2154,"type":15},"Docker","docker",{"name":9,"slug":8,"type":15},{"name":2157,"slug":2158,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":2161,"name":2161,"fn":2162,"description":2163,"org":2164,"tags":2165,"stars":20,"repoUrl":21,"updatedAt":2173},"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},[2166,2167,2170],{"name":9,"slug":8,"type":15},{"name":2168,"slug":2169,"type":15},"Quantum Computing","quantum-computing",{"name":2171,"slug":2172,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305]