[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-physical-ai-video-data-augmentation":3,"mdc-9yzqu-key":34,"related-org-nvidia-physical-ai-video-data-augmentation":4097,"related-repo-nvidia-physical-ai-video-data-augmentation":4253},{"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},"physical-ai-video-data-augmentation","run video data augmentation workflows","Use when running video data augmentation and auto-labeling workflows on OSMO: flow selection, preflight, submit-time interpolation, monitoring, and output retrieval. Trigger keywords: video data augmentation, data enrichment, auto labeling, VDA demo, OSMO workflow, pseudo labeling.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,19,20],{"name":13,"slug":14,"type":15},"Automation","automation","tag",{"name":17,"slug":18,"type":15},"Machine Learning","machine-learning",{"name":9,"slug":8,"type":15},{"name":21,"slug":22,"type":15},"Video","video",2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-14T05:29:09.19017","CC-BY-4.0 AND 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\u002Fphysical-ai-video-data-augmentation","---\nname: physical-ai-video-data-augmentation\ndescription: >-\n  Use when running video data augmentation and auto-labeling workflows on OSMO:\n  flow selection, preflight, submit-time interpolation, monitoring, and output\n  retrieval. Trigger keywords: video data augmentation, data enrichment, auto\n  labeling, VDA demo, OSMO workflow, pseudo labeling.\nlicense: CC-BY-4.0 AND Apache-2.0\nmetadata:\n  owner: NVIDIA\n  service: data\n  version: 1.0.0\n  reviewed: '2026-05-26'\n  author: NVIDIA\n  tags:\n    - physical-ai\n    - video-data-augmentation\n    - auto-labeling\n    - cosmos\n---\n\n# Physical AI Video Data Augmentation Workflow Orchestrator\n\nDefault workflow skill for VDA execution on OSMO. It owns flow selection,\npreflight, cache readiness, inference-path decisions, submit-time interpolation,\nmonitoring, and output retrieval. Component skills are consult-only.\n\n## Purpose\n\nRun the end-to-end VDA workflow safely and reproducibly from preflight to output\ndownload.\n\nDo NOT use this skill for container-internal tuning-only questions.\n\n## Prerequisites\n\nConfirm these before running preflight or any submit. Missing required secrets\nsurface as `USER_INPUT_REQUIRED:` from `scripts\u002Fpreflight_credentials.sh`.\n\n| Requirement | How it is satisfied | Used for |\n|---|---|---|\n| NGC API key (optional) | `NGC_API_KEY`, `NGC_CLI_API_KEY`, or compatible `nvapi-*` token in `NVIDIA_API_KEY`\u002F`OPENAI_API_KEY`\u002F`VLM_API_KEY`\u002F`LLM_API_KEY` | Optional for `nvcr_io` credential refresh and NGC REST scope probe; default VDA image refs are validated via workflow registry probes |\n| Hugging Face token | `HF_TOKEN` (or `HUGGING_FACE_HUB_TOKEN`), or a cached token at `~\u002F.cache\u002Fhuggingface\u002Ftoken` | Creates the OSMO `hf_token` credential; pulls gated Cosmos\u002FSeedVR weights |\n| OSMO CLI access | `osmo` on `PATH`, logged in, with a default profile and a registered DATA credential profile matching `storage_url` | Submitting\u002Fmonitoring workflows and listing\u002Fdownloading objects |\n| GPU pool | At least one `ONLINE` pool in `osmo pool list --mode free`; `POD_TEMPLATE` carries GPU toleration\u002Fselectors | Scheduling setup + worker tasks |\n\nOptional (only for the strict NGC org\u002Fteam probe): `NGC_ORG` + `NGC_TEAM`\n(or `NGC_CLI_ORG` \u002F `NGC_CLI_TEAM`). External VLM\u002FLLM endpoint keys are validated\nseparately, not by preflight.\n\nKey handling rule: `nvapi-*` tokens are first-class inputs for `nvcr_io`.\nNever reject by token prefix alone; use workflow registry probe results as\nsource of truth.\n\n## Instructions\n\n1. Select the workflow (`auto_labeling`, `augmentation_and_al`, `e2e`,\n   `e2e_super_resolution`) from user intent.\n2. Provide a tentative execution-time overview before starting run actions.\n3. Run preflight and readiness checks before submit.\n4. Derive submit-time values from the active dataset backend (never guess\n   `storage_url`).\n5. Submit the workflow with explicit interpolation values and monitor to completion.\n6. Retrieve outputs, provide side-by-side comparison evidence for augmented\n   flows, and summarize task outcomes.\n\nUse `run_script(...)` for script execution. Canonical examples:\n\n```python\nrun_script(\"bash scripts\u002Fpreflight_credentials.sh --workflow assets\u002Fconfigs\u002Fosmo\u002Faugmentation_and_al.yaml\")\nrun_script(\"python3 scripts\u002Fpre_submit_guard.py --workflow assets\u002Fconfigs\u002Fosmo\u002Fauto_labeling.yaml\")\nrun_script(\"bash scripts\u002Fprepare_demo_assets.sh \u002Fsrv\u002Fsdg\u002Fdata\u002Fvda_inputs\")\n```\n\n## Available Scripts\n\nUse script-level `--help` for exact arguments.\n\n| Script | Role |\n|---|---|\n| `scripts\u002Fpreflight_credentials.sh` | Secrets\u002Fcontrol-plane preflight and workflow image access checks |\n| `scripts\u002Fpre_submit_guard.py` | Submit-time interpolation, cache, and dataset safety checks |\n| `scripts\u002Fprepare_demo_assets.sh` | Demo video pull + flatten for default demo path |\n| `scripts\u002Fgenerate_configs.py` | Setup-time config and cookbook projection generation |\n| `scripts\u002Fcosmos_worker.sh` | Augmentation worker execution |\n| `scripts\u002Fpl_original_worker.sh` | Original-video auto-labeling worker execution |\n| `scripts\u002Fpl_augmented_worker.sh` | Augmented-video auto-labeling worker execution |\n| `scripts\u002Fosmo_barrier.py` | Multi-node barrier synchronization |\n| `scripts\u002Fstage_run_artifacts.sh` | Local mirror of full run output + input video |\n| `scripts\u002Frender_side_by_side.sh` | Side-by-side comparison render from local artifacts |\n\n## Supported Flows\n\n| Flow | OSMO YAML | Group sequence | Typical use |\n|---|---|---|---|\n| `augmentation_and_al` | `assets\u002Fconfigs\u002Fosmo\u002Faugmentation_and_al.yaml` | setup -> augmentation -> auto_labeling_augmented | Augment one or more videos, then auto-label augmented outputs |\n| `auto_labeling` | `assets\u002Fconfigs\u002Fosmo\u002Fauto_labeling.yaml` | setup -> auto_labeling | Label original videos only |\n| `e2e` | `assets\u002Fconfigs\u002Fosmo\u002Fe2e.yaml` | setup -> (auto_labeling_original + augmentation) -> auto_labeling_augmented | Throughput-first path |\n| `e2e_super_resolution` | `assets\u002Fconfigs\u002Fosmo\u002Fe2e_super_resolution.yaml` | setup -> auto_labeling_original -> augmentation -> auto_labeling_augmented | Sequential path with SR gate before augmentation |\n\nLegacy alias `assets\u002Fconfigs\u002Fosmo\u002Faugmentation_and_pl.yaml` remains for\nbackwards compatibility.\n\n### Pick the right workflow for the user's request\n\n| User intent | Workflow |\n|---|---|\n| \"Label my source videos\" \u002F \"PL-only\" \u002F \"no augmentation\" | `auto_labeling` |\n| \"Create augmented videos and label them\" | `augmentation_and_al` |\n| \"Run the full pipeline quickly\" | `e2e` |\n| \"Run full pipeline, but gate on SR-enhanced originals first\" | `e2e_super_resolution` |\n\n## Disambiguation: handle vague requests before committing\n\nDefault to autonomy: ask only when missing information blocks execution.\n\n### Autonomous defaults (do NOT ask)\n\n- If dataset source is absent, run VDA demo path (`scripts\u002Fprepare_demo_assets.sh`)\n  and continue with `dataset=vda-demo`.\n- If flow is not explicitly requested, default to `augmentation_and_al`.\n- If endpoint mode is unspecified, default to in-cluster persistent NIM reuse and\n  automatic NIM deploy\u002Frepair when unhealthy.\n- If cache is missing, run `setup_model_cache.yaml`, rerun pre-submit guard, and\n  continue automatically on success.\n- After any stage completes successfully, continue to the next stage immediately.\n  Do not pause with \"Ready when you are\" or equivalent approval prompts.\n\n### Triggers that should pause for disambiguation\n\n| Missing input | Why it matters | Ask |\n|---|---|---|\n| `USER_INPUT_REQUIRED` from preflight | Required secret is missing | Ask one concise unblock question for exactly the missing value(s) |\n| Storage backend prefix cannot be derived from the active dataset\u002Fupload root | Wrong scheme causes runtime storage auth mismatch | \"What is the backend-native root prefix for this run?\" |\n| No ONLINE GPU pool\u002Fplatform can be selected | Workflow cannot schedule setup\u002Fworkers | \"Which GPU pool\u002Fplatform should this run target?\" |\n\n### When NOT to disambiguate\n\n- Do not ask for cookbook unless user explicitly asks to change scene profile.\n- Do not offer external endpoints by default.\n- Do not ask A\u002FB cache strategy questions; default is automatic cache setup.\n- Do not ask to scale down existing NIMs; this is forbidden.\n- Do not invent, scrape, or generate random videos when input is missing.\n- Do not use non-VDA demo sources (for example Carline adaptation assets) unless\n  the user explicitly requests a different dataset.\n\n## Step 0: Select Flow and Gather Inputs\n\n### Input video policy (non-negotiable)\n\n- Always preserve user-provided video inputs (dataset URL, local path, or upload\n  folder) as first-class and preferred.\n- Never replace an explicit user video with demo assets or any other source.\n- If no video input is provided, default to VDA demo assets via\n  `scripts\u002Fprepare_demo_assets.sh` (HF dataset flow) without asking extra\n  source-selection questions.\n- If the user explicitly mentions an input video or dataset, prefer and use that\n  input instead of demo assets.\n- Use only VDA demo assets (`nvidia\u002Fvideo-data-augmentation-demo`) for the\n  default demo path.\n- Never propose arbitrary web clip downloads or placeholder videos\n  unless the user explicitly requests that behavior.\n\nCollect only missing values:\n\n1. Dataset source (prefer explicit user-provided `dataset_url` or local upload\n   folder; otherwise default to VDA demo assets and proceed).\n2. Flow (`auto_labeling`, `augmentation_and_al`, `e2e`, `e2e_super_resolution`);\n   default to `augmentation_and_al` when unspecified.\n3. OSMO `gpu_platform` for all VDA resources (auto-select an ONLINE platform\n   when unambiguous; ask only when no valid option exists).\n4. Endpoint mode (default in-cluster NIM reuse\u002Fdeploy unless explicitly\n   overridden).\n\nDo not guess `gpu_platform` (for example `microk8s`). Use the exact current\nplatform label shown by `osmo pool list --mode free` (for example `gpu`).\n\nGenerate run stamp before each submit:\n\n```bash\nSTAMP=$(cat \u002Fproc\u002Fsys\u002Fkernel\u002Frandom\u002Fuuid | cut -c1-8)\nRUN_ID=\"run-$STAMP\"\n```\n\n## Execution Time Overview (required before run)\n\nBefore running any mutating command (`osmo credential set`, NIM install\u002Frepair,\ncache workflow submit, or target VDA workflow submit), provide a short ETA\noverview to the user.\n\nKeep it concise (one short paragraph or 4-6 bullets) and include:\n\n- whether this looks like a **cold start** (NIM\u002Fcache missing) or **warm start**\n  (NIM\u002Fcache already healthy),\n- major phases with approximate durations,\n- a total expected range for the selected workflow.\n\nBaseline ranges (from observed MicroK8s + OSMO runs):\n\n| Phase | Typical duration |\n|---|---|\n| Credentials + preflight | ~1-2 min |\n| NIM deploy\u002Fdownload\u002Fwarmup (if needed) | ~10-15 min |\n| Demo assets download\u002Fupload (if demo path) | ~1-3 min |\n| Model cache population (if needed) | ~15-25 min |\n| Workflow submit + queue\u002Fstart | ~1-3 min |\n\nWorkflow runtime ranges after submit:\n\n| Flow | Typical runtime |\n|---|---|\n| `auto_labeling` | ~6-15 min |\n| `augmentation_and_al` | ~20-35 min |\n| `e2e` | ~22-40 min |\n| `e2e_super_resolution` | ~25-45 min |\n\nCold-start end-to-end runs are commonly ~45-80 min; warm-start runs are usually\n~20-45 min depending on flow and video length.\n\n## Common Preconditions (all flows)\n\n1. **Credential and control-plane preflight**\n\n   ```bash\n   bash scripts\u002Fpreflight_credentials.sh --workflow assets\u002Fconfigs\u002Fosmo\u002F\u003Cmode>.yaml\n   ```\n\n   Restricted egress:\n\n   ```bash\n   bash scripts\u002Fpreflight_credentials.sh --no-probe --workflow assets\u002Fconfigs\u002Fosmo\u002F\u003Cmode>.yaml\n   ```\n\n   Preflight does not require a workload-local `.env`. Runtime interpolation is\n   driven by submit-time values (`dataset`, `run_id`, `gpu_platform`, `video`,\n   `storage_url`, `skills_dir`) supplied in one `--set-string` list.\n\n   Passing `--workflow` validates pull access for the active workflow image refs\n   (`workflow.groups[].tasks[].image`) using anonymous bearer access with\n   credential fallback when provided.\n   If replacement NGC\u002FHF secrets are provided in env, preflight refreshes\n   existing `nvcr_io` \u002F `hf_token` automatically when present. Use `--refresh` to force\n   overwrite even when no new env secrets were supplied:\n\n   ```bash\n   bash scripts\u002Fpreflight_credentials.sh --workflow assets\u002Fconfigs\u002Fosmo\u002F\u003Cmode>.yaml --refresh\n   ```\n\n   If output contains `USER_INPUT_REQUIRED:`, ask one concise unblock question\n   and stop.\n\n   On workflow image `401\u002F403`, report registry access failure after probe\n   checks on the listed image refs; do not claim a key family (for example\n   `nvapi-*`) is categorically unsupported.\n\n2. **Storage interpolation policy**\n\n   `storage_url` must be derived from the actual dataset\u002Fupload backend for the\n   current run.\n\n   ```text\n   dataset_url=azure:\u002F\u002Fstoriondevxah69\u002Fosmo-workflows\u002Fdatasets\u002Fvda-demo\n   storage_url=azure:\u002F\u002Fstoriondevxah69\u002Fosmo-workflows\n   dataset=vda-demo\n   ```\n\n   Never silently default to stale `s3:\u002F\u002F` values on non-S3 backends.\n\n3. **Inference policy (non-negotiable)**\n\n   - Reuse healthy in-cluster persistent NIM endpoints by default.\n   - If missing\u002Funhealthy, deploy automatically — this is a prerequisite, not a\n     user decision. Do NOT pause to ask; run the install with the VDA allow-list:\n\n   ```bash\n   export NIM_SERVICES=\"qwen3-vl qwen25-14b\"\n   skills\u002Fphysical-ai-infrastructure-setup-and-resilient-scaling\u002Fcomponents\u002Finference-nim-operator\u002Fscripts\u002Finstall.sh\n   ```\n\n   - See `references\u002Fnim\u002FREADME.md` for full endpoint docs and health checks.\n   - External endpoints are opt-in only (explicit request or explicit URLs); only\n     then skip the in-cluster deploy.\n   - Never infer external mode from credential presence.\n   - Never scale down\u002Fdelete existing NIMs to free GPUs.\n\n4. **Readiness guard**\n\n   ```bash\n   osmo pool list --mode free\n   osmo config show POD_TEMPLATE\n   python3 scripts\u002Fpre_submit_guard.py --workflow assets\u002Fconfigs\u002Fosmo\u002F\u003Cmode>.yaml\n   ```\n\n5. **Cache auto-remediation**\n\n   If `pre_submit_guard.py` reports cache failure, default action is to run:\n\n   ```bash\n   osmo workflow submit assets\u002Fconfigs\u002Fosmo\u002Fsetup_model_cache.yaml \\\n     --set-string storage_url=\u003Cbackend-prefix> path=data\n   ```\n\n   Then rerun `pre_submit_guard.py` and submit the target VDA flow only after it\n   passes. Ask user only when backend\u002Fprefix is ambiguous or cache setup fails.\n\n6. **Scheduling policy**\n\n   VDA templates schedule setup and workers on `gpu_platform` (no `system` pool\n   dependency for user workloads).\n\n## Submit (all flows)\n\nEvery flow uses the same submit shape; only the workflow YAML changes. Choose the\nYAML for the requested flow, then run the command below. Full per-flow walkthroughs\n(stage matrix and flow details) live in the linked references.\n\n| Flow | Workflow YAML | Walkthrough |\n|---|---|---|\n| Augmentation + auto-labeling | `assets\u002Fconfigs\u002Fosmo\u002Faugmentation_and_al.yaml` | `references\u002Fflows\u002Faugmentation_and_al.md` |\n| Auto-labeling only | `assets\u002Fconfigs\u002Fosmo\u002Fauto_labeling.yaml` | `references\u002Fflows\u002Fauto_labeling.md` |\n| E2E (parallel) | `assets\u002Fconfigs\u002Fosmo\u002Fe2e.yaml` | `references\u002Fflows\u002Fe2e.md` |\n| E2E (super-resolution gated) | `assets\u002Fconfigs\u002Fosmo\u002Fe2e_super_resolution.yaml` | `references\u002Fflows\u002Fe2e_super_resolution.md` |\n\n```bash\nSKILLS_DIR=\"$(cd \"$(git rev-parse --show-toplevel)\u002Fskills\u002Fphysical-ai-video-data-augmentation\" && pwd)\"\nSTAMP=$(cat \u002Fproc\u002Fsys\u002Fkernel\u002Frandom\u002Fuuid | cut -c1-8)\nosmo workflow submit assets\u002Fconfigs\u002Fosmo\u002F\u003Cflow>.yaml \\\n  --pool \u003Cpool> \\\n  --set-string \\\n    dataset=\u003Cdataset> \\\n    run_id=run-$STAMP \\\n    storage_url=\u003Cbackend-prefix> \\\n    gpu_platform=\u003Cgpu-platform> \\\n    video=\u003Cvideo-stem> \\\n    cosmos_model_cache_url=\u003Cbackend-prefix>\u002Fdata\u002Fmodels\u002Fcosmos_transfer \\\n    auto_labeling_model_cache_url=\u003Cbackend-prefix>\u002Fdata\u002Fmodels\u002Fauto_labeling \\\n    skills_dir=\"$SKILLS_DIR\"\n```\n\nCompatibility note:\n- Use exactly one `--set-string` flag and pass all the key\u002Fvalue pairs after it.\n- Do not repeat `--set`\u002F`--set-string` flags in the same command; some OSMO builds\n  only honor the last occurrence.\n- Do not mix `--set` and `--set-string` in one submit command.\n- Pass explicit `*_model_cache_url` values to avoid nested-template interpolation\n  differences across OSMO environments.\n- Do not brute-force permutations of flags. Use this shape directly.\n\nCommon optional overrides (append key\u002Fvalue pairs to the same `--set-string` list):\n\n```bash\ncookbook=\u003Cscene_profile> \\\nvlm_url=\u003Copenai_base_url> \\\nllm_url=\u003Copenai_base_url> \\\ncosmos_model_cache_url=\u003Curl> \\\nauto_labeling_model_cache_url=\u003Curl>\n```\n\nThe auto-labeling-only flow has no augmentation stage, so it omits\n`cosmos_model_cache_url` at runtime; passing it is harmless and keeps one submit\nshape across flows.\n\n## OSMO Monitoring\n\n```bash\n# Workflow status + task states\nosmo workflow query \u003Cworkflow_id> --format-type json \\\n  | jq '{status, tasks: [.groups[].tasks[] | {name, status, exit_code}]}'\n\n# Logs for a specific task\nosmo workflow logs \u003Cworkflow_id> --task \u003Ctask_name> -n 200\n\n# Output retrieval\nosmo data list --no-pager \u003Coutput_url>\nosmo data download \u003Coutput_url> \u003Clocal_dir>\u002F\n```\n\nFor completion artifacts, always mirror the full run output into workspace:\n\n```bash\nROOT=\"$(git rev-parse --show-toplevel)\"\nRUN_LOCAL_DIR=\"$ROOT\u002Fmedia\u002Fvda\u002Fruns\u002F\u003Crun_id>\"\nmkdir -p \"$RUN_LOCAL_DIR\"\nosmo data download \"\u003Cstorage_url>\u002Fdatasets\u002F\u003Cdataset>-outputs\u002F\u003Crun_id>\u002F\" \"$RUN_LOCAL_DIR\u002F\"\n```\n\nFor runs expected to exceed two minutes, send heartbeat updates at least every\ntwo minutes. For media evidence, emit one standalone `MEDIA:\u003Cabsolute-path>`\nline per message bubble.\n\nExecution continuity requirement:\n\n- Heartbeats must report progress while continuing work; they are status updates,\n  not permission prompts.\n- Do not stop between green stages waiting for approval.\n- Pause only on blocking failures or explicit user stop\u002Fredirect.\n- If submit fails on interpolation, rerun once with the same canonical single-flag\n  shape and corrected values; do not loop through ad-hoc flag experiments.\n\nMEDIA formatting is strict:\n\n- Emit exactly one line: `MEDIA:\u002Fabsolute\u002Fpath\u002Fto\u002Ffile.mp4`\n- Keep `MEDIA:` contiguous on a single line (never split across lines).\n- No extra text in the same bubble.\n- No code fences, bullets, or quotes around the directive.\n- If render fails: retry once from a stable workspace path, then emit PNG fallback.\n\n## Post-Run Comparison Evidence (required for augmented flows)\n\nApplies to `augmentation_and_al`, `e2e`, and `e2e_super_resolution` after a\nsuccessful run.\n\nRequired completion output (do not stop at raw output URLs):\n\n1. Stage full outputs + input video into workspace-local path:\n\n   ```bash\n   bash scripts\u002Fstage_run_artifacts.sh \\\n     --storage-url \u003Cstorage_url> --dataset \u003Cdataset> --run-id \u003Crun_id> --video \u003Cvideo>\n   ```\n\n2. Render side-by-side from that local run copy:\n\n   ```bash\n   bash scripts\u002Frender_side_by_side.sh \\\n     --run-local-dir \"\u003Crepo>\u002Fmedia\u002Fvda\u002Fruns\u002F\u003Crun_id>\" --dataset \u003Cdataset> --video \u003Cvideo>\n   ```\n\n3. Emit MEDIA from the local run copy and include:\n   - augmentation summary from `\u003Crun_local_dir>\u002Fsetup_b0\u002Fconfigs\u002Fmanifest.yaml`\n     (`sampled_vars` for `\u003Cvideo>_aug0`)\n   - auto-labeling summary from `\u003Crun_local_dir>\u002Foutputs\u002Fpseudo_labeled_augmented\u002F\u003Cvideo>_aug0`\n   - for `e2e` \u002F `e2e_super_resolution`, original-label summary from\n     `\u003Crun_local_dir>\u002Foutputs\u002Fpseudo_labeled\u002F\u003Cvideo>`\n\nIf `ffmpeg` is unavailable, emit input and augmented MEDIA from the same local\nrun copy and still provide augmentation + auto-labeling summaries.\n\nFor demo runs (no user video provided), explicitly state that input came from\n`nvidia\u002Fvideo-data-augmentation-demo`.\n\n## Supporting files\n\nUse these canonical locations:\n\n- Workflows: `assets\u002Fconfigs\u002Fosmo\u002F*.yaml`\n- Runtime scripts: `scripts\u002F*.sh`, `scripts\u002F*.py`\n- Flow walkthroughs: `references\u002Fflows\u002F*.md`\n- Setup and triage: `references\u002Fsetup.md`, `references\u002Ftroubleshooting.md`\n- Images and endpoint policy: `references\u002Fcontainer-images.md`, `references\u002Fnim\u002FREADME.md`\n- Cookbook tuning: `assets\u002Fcookbooks\u002FTUNING_GUIDE.md`\n",{"data":35,"body":45},{"name":4,"description":6,"license":26,"metadata":36},{"owner":9,"service":37,"version":38,"reviewed":39,"author":9,"tags":40},"data","1.0.0","2026-05-26",[41,42,43,44],"physical-ai","video-data-augmentation","auto-labeling","cosmos",{"type":46,"children":47},"root",[48,57,63,70,75,80,86,108,343,380,399,405,477,490,530,536,549,740,746,898,911,918,1004,1010,1015,1021,1078,1084,1171,1177,1210,1216,1222,1270,1275,1346,1379,1384,1473,1479,1492,1497,1530,1535,1621,1626,1711,1716,1722,2406,2412,2417,2543,2966,2971,3042,3054,3119,3132,3138,3421,3426,3562,3575,3580,3603,3608,3650,3656,3681,3686,3968,3980,3991,3997,4002,4091],{"type":49,"tag":50,"props":51,"children":53},"element","h1",{"id":52},"physical-ai-video-data-augmentation-workflow-orchestrator",[54],{"type":55,"value":56},"text","Physical AI Video Data Augmentation Workflow Orchestrator",{"type":49,"tag":58,"props":59,"children":60},"p",{},[61],{"type":55,"value":62},"Default workflow skill for VDA execution on OSMO. It owns flow selection,\npreflight, cache readiness, inference-path decisions, submit-time interpolation,\nmonitoring, and output retrieval. Component skills are consult-only.",{"type":49,"tag":64,"props":65,"children":67},"h2",{"id":66},"purpose",[68],{"type":55,"value":69},"Purpose",{"type":49,"tag":58,"props":71,"children":72},{},[73],{"type":55,"value":74},"Run the end-to-end VDA workflow safely and reproducibly from preflight to output\ndownload.",{"type":49,"tag":58,"props":76,"children":77},{},[78],{"type":55,"value":79},"Do NOT use this skill for container-internal tuning-only questions.",{"type":49,"tag":64,"props":81,"children":83},{"id":82},"prerequisites",[84],{"type":55,"value":85},"Prerequisites",{"type":49,"tag":58,"props":87,"children":88},{},[89,91,98,100,106],{"type":55,"value":90},"Confirm these before running preflight or any submit. Missing required secrets\nsurface as ",{"type":49,"tag":92,"props":93,"children":95},"code",{"className":94},[],[96],{"type":55,"value":97},"USER_INPUT_REQUIRED:",{"type":55,"value":99}," from ",{"type":49,"tag":92,"props":101,"children":103},{"className":102},[],[104],{"type":55,"value":105},"scripts\u002Fpreflight_credentials.sh",{"type":55,"value":107},".",{"type":49,"tag":109,"props":110,"children":111},"table",{},[112,136],{"type":49,"tag":113,"props":114,"children":115},"thead",{},[116],{"type":49,"tag":117,"props":118,"children":119},"tr",{},[120,126,131],{"type":49,"tag":121,"props":122,"children":123},"th",{},[124],{"type":55,"value":125},"Requirement",{"type":49,"tag":121,"props":127,"children":128},{},[129],{"type":55,"value":130},"How it is satisfied",{"type":49,"tag":121,"props":132,"children":133},{},[134],{"type":55,"value":135},"Used for",{"type":49,"tag":137,"props":138,"children":139},"tbody",{},[140,217,263,301],{"type":49,"tag":117,"props":141,"children":142},{},[143,149,204],{"type":49,"tag":144,"props":145,"children":146},"td",{},[147],{"type":55,"value":148},"NGC API key (optional)",{"type":49,"tag":144,"props":150,"children":151},{},[152,158,160,166,168,174,176,182,184,190,191,197,198],{"type":49,"tag":92,"props":153,"children":155},{"className":154},[],[156],{"type":55,"value":157},"NGC_API_KEY",{"type":55,"value":159},", ",{"type":49,"tag":92,"props":161,"children":163},{"className":162},[],[164],{"type":55,"value":165},"NGC_CLI_API_KEY",{"type":55,"value":167},", or compatible ",{"type":49,"tag":92,"props":169,"children":171},{"className":170},[],[172],{"type":55,"value":173},"nvapi-*",{"type":55,"value":175}," token in ",{"type":49,"tag":92,"props":177,"children":179},{"className":178},[],[180],{"type":55,"value":181},"NVIDIA_API_KEY",{"type":55,"value":183},"\u002F",{"type":49,"tag":92,"props":185,"children":187},{"className":186},[],[188],{"type":55,"value":189},"OPENAI_API_KEY",{"type":55,"value":183},{"type":49,"tag":92,"props":192,"children":194},{"className":193},[],[195],{"type":55,"value":196},"VLM_API_KEY",{"type":55,"value":183},{"type":49,"tag":92,"props":199,"children":201},{"className":200},[],[202],{"type":55,"value":203},"LLM_API_KEY",{"type":49,"tag":144,"props":205,"children":206},{},[207,209,215],{"type":55,"value":208},"Optional for ",{"type":49,"tag":92,"props":210,"children":212},{"className":211},[],[213],{"type":55,"value":214},"nvcr_io",{"type":55,"value":216}," credential refresh and NGC REST scope probe; default VDA image refs are validated via workflow registry probes",{"type":49,"tag":117,"props":218,"children":219},{},[220,225,250],{"type":49,"tag":144,"props":221,"children":222},{},[223],{"type":55,"value":224},"Hugging Face token",{"type":49,"tag":144,"props":226,"children":227},{},[228,234,236,242,244],{"type":49,"tag":92,"props":229,"children":231},{"className":230},[],[232],{"type":55,"value":233},"HF_TOKEN",{"type":55,"value":235}," (or ",{"type":49,"tag":92,"props":237,"children":239},{"className":238},[],[240],{"type":55,"value":241},"HUGGING_FACE_HUB_TOKEN",{"type":55,"value":243},"), or a cached token at ",{"type":49,"tag":92,"props":245,"children":247},{"className":246},[],[248],{"type":55,"value":249},"~\u002F.cache\u002Fhuggingface\u002Ftoken",{"type":49,"tag":144,"props":251,"children":252},{},[253,255,261],{"type":55,"value":254},"Creates the OSMO ",{"type":49,"tag":92,"props":256,"children":258},{"className":257},[],[259],{"type":55,"value":260},"hf_token",{"type":55,"value":262}," credential; pulls gated Cosmos\u002FSeedVR weights",{"type":49,"tag":117,"props":264,"children":265},{},[266,271,296],{"type":49,"tag":144,"props":267,"children":268},{},[269],{"type":55,"value":270},"OSMO CLI access",{"type":49,"tag":144,"props":272,"children":273},{},[274,280,282,288,290],{"type":49,"tag":92,"props":275,"children":277},{"className":276},[],[278],{"type":55,"value":279},"osmo",{"type":55,"value":281}," on ",{"type":49,"tag":92,"props":283,"children":285},{"className":284},[],[286],{"type":55,"value":287},"PATH",{"type":55,"value":289},", logged in, with a default profile and a registered DATA credential profile matching ",{"type":49,"tag":92,"props":291,"children":293},{"className":292},[],[294],{"type":55,"value":295},"storage_url",{"type":49,"tag":144,"props":297,"children":298},{},[299],{"type":55,"value":300},"Submitting\u002Fmonitoring workflows and listing\u002Fdownloading objects",{"type":49,"tag":117,"props":302,"children":303},{},[304,309,338],{"type":49,"tag":144,"props":305,"children":306},{},[307],{"type":55,"value":308},"GPU pool",{"type":49,"tag":144,"props":310,"children":311},{},[312,314,320,322,328,330,336],{"type":55,"value":313},"At least one ",{"type":49,"tag":92,"props":315,"children":317},{"className":316},[],[318],{"type":55,"value":319},"ONLINE",{"type":55,"value":321}," pool in ",{"type":49,"tag":92,"props":323,"children":325},{"className":324},[],[326],{"type":55,"value":327},"osmo pool list --mode free",{"type":55,"value":329},"; ",{"type":49,"tag":92,"props":331,"children":333},{"className":332},[],[334],{"type":55,"value":335},"POD_TEMPLATE",{"type":55,"value":337}," carries GPU toleration\u002Fselectors",{"type":49,"tag":144,"props":339,"children":340},{},[341],{"type":55,"value":342},"Scheduling setup + worker tasks",{"type":49,"tag":58,"props":344,"children":345},{},[346,348,354,356,362,364,370,372,378],{"type":55,"value":347},"Optional (only for the strict NGC org\u002Fteam probe): ",{"type":49,"tag":92,"props":349,"children":351},{"className":350},[],[352],{"type":55,"value":353},"NGC_ORG",{"type":55,"value":355}," + ",{"type":49,"tag":92,"props":357,"children":359},{"className":358},[],[360],{"type":55,"value":361},"NGC_TEAM",{"type":55,"value":363},"\n(or ",{"type":49,"tag":92,"props":365,"children":367},{"className":366},[],[368],{"type":55,"value":369},"NGC_CLI_ORG",{"type":55,"value":371}," \u002F ",{"type":49,"tag":92,"props":373,"children":375},{"className":374},[],[376],{"type":55,"value":377},"NGC_CLI_TEAM",{"type":55,"value":379},"). External VLM\u002FLLM endpoint keys are validated\nseparately, not by preflight.",{"type":49,"tag":58,"props":381,"children":382},{},[383,385,390,392,397],{"type":55,"value":384},"Key handling rule: ",{"type":49,"tag":92,"props":386,"children":388},{"className":387},[],[389],{"type":55,"value":173},{"type":55,"value":391}," tokens are first-class inputs for ",{"type":49,"tag":92,"props":393,"children":395},{"className":394},[],[396],{"type":55,"value":214},{"type":55,"value":398},".\nNever reject by token prefix alone; use workflow registry probe results as\nsource of truth.",{"type":49,"tag":64,"props":400,"children":402},{"id":401},"instructions",[403],{"type":55,"value":404},"Instructions",{"type":49,"tag":406,"props":407,"children":408},"ol",{},[409,445,450,455,467,472],{"type":49,"tag":410,"props":411,"children":412},"li",{},[413,415,421,422,428,429,435,437,443],{"type":55,"value":414},"Select the workflow (",{"type":49,"tag":92,"props":416,"children":418},{"className":417},[],[419],{"type":55,"value":420},"auto_labeling",{"type":55,"value":159},{"type":49,"tag":92,"props":423,"children":425},{"className":424},[],[426],{"type":55,"value":427},"augmentation_and_al",{"type":55,"value":159},{"type":49,"tag":92,"props":430,"children":432},{"className":431},[],[433],{"type":55,"value":434},"e2e",{"type":55,"value":436},",\n",{"type":49,"tag":92,"props":438,"children":440},{"className":439},[],[441],{"type":55,"value":442},"e2e_super_resolution",{"type":55,"value":444},") from user intent.",{"type":49,"tag":410,"props":446,"children":447},{},[448],{"type":55,"value":449},"Provide a tentative execution-time overview before starting run actions.",{"type":49,"tag":410,"props":451,"children":452},{},[453],{"type":55,"value":454},"Run preflight and readiness checks before submit.",{"type":49,"tag":410,"props":456,"children":457},{},[458,460,465],{"type":55,"value":459},"Derive submit-time values from the active dataset backend (never guess\n",{"type":49,"tag":92,"props":461,"children":463},{"className":462},[],[464],{"type":55,"value":295},{"type":55,"value":466},").",{"type":49,"tag":410,"props":468,"children":469},{},[470],{"type":55,"value":471},"Submit the workflow with explicit interpolation values and monitor to completion.",{"type":49,"tag":410,"props":473,"children":474},{},[475],{"type":55,"value":476},"Retrieve outputs, provide side-by-side comparison evidence for augmented\nflows, and summarize task outcomes.",{"type":49,"tag":58,"props":478,"children":479},{},[480,482,488],{"type":55,"value":481},"Use ",{"type":49,"tag":92,"props":483,"children":485},{"className":484},[],[486],{"type":55,"value":487},"run_script(...)",{"type":55,"value":489}," for script execution. Canonical examples:",{"type":49,"tag":491,"props":492,"children":497},"pre",{"className":493,"code":494,"language":495,"meta":496,"style":496},"language-python shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","run_script(\"bash scripts\u002Fpreflight_credentials.sh --workflow assets\u002Fconfigs\u002Fosmo\u002Faugmentation_and_al.yaml\")\nrun_script(\"python3 scripts\u002Fpre_submit_guard.py --workflow assets\u002Fconfigs\u002Fosmo\u002Fauto_labeling.yaml\")\nrun_script(\"bash scripts\u002Fprepare_demo_assets.sh \u002Fsrv\u002Fsdg\u002Fdata\u002Fvda_inputs\")\n","python","",[498],{"type":49,"tag":92,"props":499,"children":500},{"__ignoreMap":496},[501,512,521],{"type":49,"tag":502,"props":503,"children":506},"span",{"class":504,"line":505},"line",1,[507],{"type":49,"tag":502,"props":508,"children":509},{},[510],{"type":55,"value":511},"run_script(\"bash scripts\u002Fpreflight_credentials.sh --workflow assets\u002Fconfigs\u002Fosmo\u002Faugmentation_and_al.yaml\")\n",{"type":49,"tag":502,"props":513,"children":515},{"class":504,"line":514},2,[516],{"type":49,"tag":502,"props":517,"children":518},{},[519],{"type":55,"value":520},"run_script(\"python3 scripts\u002Fpre_submit_guard.py --workflow assets\u002Fconfigs\u002Fosmo\u002Fauto_labeling.yaml\")\n",{"type":49,"tag":502,"props":522,"children":524},{"class":504,"line":523},3,[525],{"type":49,"tag":502,"props":526,"children":527},{},[528],{"type":55,"value":529},"run_script(\"bash scripts\u002Fprepare_demo_assets.sh \u002Fsrv\u002Fsdg\u002Fdata\u002Fvda_inputs\")\n",{"type":49,"tag":64,"props":531,"children":533},{"id":532},"available-scripts",[534],{"type":55,"value":535},"Available Scripts",{"type":49,"tag":58,"props":537,"children":538},{},[539,541,547],{"type":55,"value":540},"Use script-level ",{"type":49,"tag":92,"props":542,"children":544},{"className":543},[],[545],{"type":55,"value":546},"--help",{"type":55,"value":548}," for exact arguments.",{"type":49,"tag":109,"props":550,"children":551},{},[552,568],{"type":49,"tag":113,"props":553,"children":554},{},[555],{"type":49,"tag":117,"props":556,"children":557},{},[558,563],{"type":49,"tag":121,"props":559,"children":560},{},[561],{"type":55,"value":562},"Script",{"type":49,"tag":121,"props":564,"children":565},{},[566],{"type":55,"value":567},"Role",{"type":49,"tag":137,"props":569,"children":570},{},[571,587,604,621,638,655,672,689,706,723],{"type":49,"tag":117,"props":572,"children":573},{},[574,582],{"type":49,"tag":144,"props":575,"children":576},{},[577],{"type":49,"tag":92,"props":578,"children":580},{"className":579},[],[581],{"type":55,"value":105},{"type":49,"tag":144,"props":583,"children":584},{},[585],{"type":55,"value":586},"Secrets\u002Fcontrol-plane preflight and workflow image access checks",{"type":49,"tag":117,"props":588,"children":589},{},[590,599],{"type":49,"tag":144,"props":591,"children":592},{},[593],{"type":49,"tag":92,"props":594,"children":596},{"className":595},[],[597],{"type":55,"value":598},"scripts\u002Fpre_submit_guard.py",{"type":49,"tag":144,"props":600,"children":601},{},[602],{"type":55,"value":603},"Submit-time interpolation, cache, and dataset safety checks",{"type":49,"tag":117,"props":605,"children":606},{},[607,616],{"type":49,"tag":144,"props":608,"children":609},{},[610],{"type":49,"tag":92,"props":611,"children":613},{"className":612},[],[614],{"type":55,"value":615},"scripts\u002Fprepare_demo_assets.sh",{"type":49,"tag":144,"props":617,"children":618},{},[619],{"type":55,"value":620},"Demo video pull + flatten for default demo path",{"type":49,"tag":117,"props":622,"children":623},{},[624,633],{"type":49,"tag":144,"props":625,"children":626},{},[627],{"type":49,"tag":92,"props":628,"children":630},{"className":629},[],[631],{"type":55,"value":632},"scripts\u002Fgenerate_configs.py",{"type":49,"tag":144,"props":634,"children":635},{},[636],{"type":55,"value":637},"Setup-time config and cookbook projection generation",{"type":49,"tag":117,"props":639,"children":640},{},[641,650],{"type":49,"tag":144,"props":642,"children":643},{},[644],{"type":49,"tag":92,"props":645,"children":647},{"className":646},[],[648],{"type":55,"value":649},"scripts\u002Fcosmos_worker.sh",{"type":49,"tag":144,"props":651,"children":652},{},[653],{"type":55,"value":654},"Augmentation worker execution",{"type":49,"tag":117,"props":656,"children":657},{},[658,667],{"type":49,"tag":144,"props":659,"children":660},{},[661],{"type":49,"tag":92,"props":662,"children":664},{"className":663},[],[665],{"type":55,"value":666},"scripts\u002Fpl_original_worker.sh",{"type":49,"tag":144,"props":668,"children":669},{},[670],{"type":55,"value":671},"Original-video auto-labeling worker execution",{"type":49,"tag":117,"props":673,"children":674},{},[675,684],{"type":49,"tag":144,"props":676,"children":677},{},[678],{"type":49,"tag":92,"props":679,"children":681},{"className":680},[],[682],{"type":55,"value":683},"scripts\u002Fpl_augmented_worker.sh",{"type":49,"tag":144,"props":685,"children":686},{},[687],{"type":55,"value":688},"Augmented-video auto-labeling worker execution",{"type":49,"tag":117,"props":690,"children":691},{},[692,701],{"type":49,"tag":144,"props":693,"children":694},{},[695],{"type":49,"tag":92,"props":696,"children":698},{"className":697},[],[699],{"type":55,"value":700},"scripts\u002Fosmo_barrier.py",{"type":49,"tag":144,"props":702,"children":703},{},[704],{"type":55,"value":705},"Multi-node barrier synchronization",{"type":49,"tag":117,"props":707,"children":708},{},[709,718],{"type":49,"tag":144,"props":710,"children":711},{},[712],{"type":49,"tag":92,"props":713,"children":715},{"className":714},[],[716],{"type":55,"value":717},"scripts\u002Fstage_run_artifacts.sh",{"type":49,"tag":144,"props":719,"children":720},{},[721],{"type":55,"value":722},"Local mirror of full run output + input video",{"type":49,"tag":117,"props":724,"children":725},{},[726,735],{"type":49,"tag":144,"props":727,"children":728},{},[729],{"type":49,"tag":92,"props":730,"children":732},{"className":731},[],[733],{"type":55,"value":734},"scripts\u002Frender_side_by_side.sh",{"type":49,"tag":144,"props":736,"children":737},{},[738],{"type":55,"value":739},"Side-by-side comparison render from local artifacts",{"type":49,"tag":64,"props":741,"children":743},{"id":742},"supported-flows",[744],{"type":55,"value":745},"Supported Flows",{"type":49,"tag":109,"props":747,"children":748},{},[749,775],{"type":49,"tag":113,"props":750,"children":751},{},[752],{"type":49,"tag":117,"props":753,"children":754},{},[755,760,765,770],{"type":49,"tag":121,"props":756,"children":757},{},[758],{"type":55,"value":759},"Flow",{"type":49,"tag":121,"props":761,"children":762},{},[763],{"type":55,"value":764},"OSMO YAML",{"type":49,"tag":121,"props":766,"children":767},{},[768],{"type":55,"value":769},"Group sequence",{"type":49,"tag":121,"props":771,"children":772},{},[773],{"type":55,"value":774},"Typical use",{"type":49,"tag":137,"props":776,"children":777},{},[778,808,838,868],{"type":49,"tag":117,"props":779,"children":780},{},[781,789,798,803],{"type":49,"tag":144,"props":782,"children":783},{},[784],{"type":49,"tag":92,"props":785,"children":787},{"className":786},[],[788],{"type":55,"value":427},{"type":49,"tag":144,"props":790,"children":791},{},[792],{"type":49,"tag":92,"props":793,"children":795},{"className":794},[],[796],{"type":55,"value":797},"assets\u002Fconfigs\u002Fosmo\u002Faugmentation_and_al.yaml",{"type":49,"tag":144,"props":799,"children":800},{},[801],{"type":55,"value":802},"setup -> augmentation -> auto_labeling_augmented",{"type":49,"tag":144,"props":804,"children":805},{},[806],{"type":55,"value":807},"Augment one or more videos, then auto-label augmented outputs",{"type":49,"tag":117,"props":809,"children":810},{},[811,819,828,833],{"type":49,"tag":144,"props":812,"children":813},{},[814],{"type":49,"tag":92,"props":815,"children":817},{"className":816},[],[818],{"type":55,"value":420},{"type":49,"tag":144,"props":820,"children":821},{},[822],{"type":49,"tag":92,"props":823,"children":825},{"className":824},[],[826],{"type":55,"value":827},"assets\u002Fconfigs\u002Fosmo\u002Fauto_labeling.yaml",{"type":49,"tag":144,"props":829,"children":830},{},[831],{"type":55,"value":832},"setup -> auto_labeling",{"type":49,"tag":144,"props":834,"children":835},{},[836],{"type":55,"value":837},"Label original videos only",{"type":49,"tag":117,"props":839,"children":840},{},[841,849,858,863],{"type":49,"tag":144,"props":842,"children":843},{},[844],{"type":49,"tag":92,"props":845,"children":847},{"className":846},[],[848],{"type":55,"value":434},{"type":49,"tag":144,"props":850,"children":851},{},[852],{"type":49,"tag":92,"props":853,"children":855},{"className":854},[],[856],{"type":55,"value":857},"assets\u002Fconfigs\u002Fosmo\u002Fe2e.yaml",{"type":49,"tag":144,"props":859,"children":860},{},[861],{"type":55,"value":862},"setup -> (auto_labeling_original + augmentation) -> auto_labeling_augmented",{"type":49,"tag":144,"props":864,"children":865},{},[866],{"type":55,"value":867},"Throughput-first path",{"type":49,"tag":117,"props":869,"children":870},{},[871,879,888,893],{"type":49,"tag":144,"props":872,"children":873},{},[874],{"type":49,"tag":92,"props":875,"children":877},{"className":876},[],[878],{"type":55,"value":442},{"type":49,"tag":144,"props":880,"children":881},{},[882],{"type":49,"tag":92,"props":883,"children":885},{"className":884},[],[886],{"type":55,"value":887},"assets\u002Fconfigs\u002Fosmo\u002Fe2e_super_resolution.yaml",{"type":49,"tag":144,"props":889,"children":890},{},[891],{"type":55,"value":892},"setup -> auto_labeling_original -> augmentation -> auto_labeling_augmented",{"type":49,"tag":144,"props":894,"children":895},{},[896],{"type":55,"value":897},"Sequential path with SR gate before augmentation",{"type":49,"tag":58,"props":899,"children":900},{},[901,903,909],{"type":55,"value":902},"Legacy alias ",{"type":49,"tag":92,"props":904,"children":906},{"className":905},[],[907],{"type":55,"value":908},"assets\u002Fconfigs\u002Fosmo\u002Faugmentation_and_pl.yaml",{"type":55,"value":910}," remains for\nbackwards compatibility.",{"type":49,"tag":912,"props":913,"children":915},"h3",{"id":914},"pick-the-right-workflow-for-the-users-request",[916],{"type":55,"value":917},"Pick the right workflow for the user's request",{"type":49,"tag":109,"props":919,"children":920},{},[921,937],{"type":49,"tag":113,"props":922,"children":923},{},[924],{"type":49,"tag":117,"props":925,"children":926},{},[927,932],{"type":49,"tag":121,"props":928,"children":929},{},[930],{"type":55,"value":931},"User intent",{"type":49,"tag":121,"props":933,"children":934},{},[935],{"type":55,"value":936},"Workflow",{"type":49,"tag":137,"props":938,"children":939},{},[940,956,972,988],{"type":49,"tag":117,"props":941,"children":942},{},[943,948],{"type":49,"tag":144,"props":944,"children":945},{},[946],{"type":55,"value":947},"\"Label my source videos\" \u002F \"PL-only\" \u002F \"no augmentation\"",{"type":49,"tag":144,"props":949,"children":950},{},[951],{"type":49,"tag":92,"props":952,"children":954},{"className":953},[],[955],{"type":55,"value":420},{"type":49,"tag":117,"props":957,"children":958},{},[959,964],{"type":49,"tag":144,"props":960,"children":961},{},[962],{"type":55,"value":963},"\"Create augmented videos and label them\"",{"type":49,"tag":144,"props":965,"children":966},{},[967],{"type":49,"tag":92,"props":968,"children":970},{"className":969},[],[971],{"type":55,"value":427},{"type":49,"tag":117,"props":973,"children":974},{},[975,980],{"type":49,"tag":144,"props":976,"children":977},{},[978],{"type":55,"value":979},"\"Run the full pipeline quickly\"",{"type":49,"tag":144,"props":981,"children":982},{},[983],{"type":49,"tag":92,"props":984,"children":986},{"className":985},[],[987],{"type":55,"value":434},{"type":49,"tag":117,"props":989,"children":990},{},[991,996],{"type":49,"tag":144,"props":992,"children":993},{},[994],{"type":55,"value":995},"\"Run full pipeline, but gate on SR-enhanced originals first\"",{"type":49,"tag":144,"props":997,"children":998},{},[999],{"type":49,"tag":92,"props":1000,"children":1002},{"className":1001},[],[1003],{"type":55,"value":442},{"type":49,"tag":64,"props":1005,"children":1007},{"id":1006},"disambiguation-handle-vague-requests-before-committing",[1008],{"type":55,"value":1009},"Disambiguation: handle vague requests before committing",{"type":49,"tag":58,"props":1011,"children":1012},{},[1013],{"type":55,"value":1014},"Default to autonomy: ask only when missing information blocks execution.",{"type":49,"tag":912,"props":1016,"children":1018},{"id":1017},"autonomous-defaults-do-not-ask",[1019],{"type":55,"value":1020},"Autonomous defaults (do NOT ask)",{"type":49,"tag":1022,"props":1023,"children":1024},"ul",{},[1025,1044,1055,1060,1073],{"type":49,"tag":410,"props":1026,"children":1027},{},[1028,1030,1035,1037,1043],{"type":55,"value":1029},"If dataset source is absent, run VDA demo path (",{"type":49,"tag":92,"props":1031,"children":1033},{"className":1032},[],[1034],{"type":55,"value":615},{"type":55,"value":1036},")\nand continue with ",{"type":49,"tag":92,"props":1038,"children":1040},{"className":1039},[],[1041],{"type":55,"value":1042},"dataset=vda-demo",{"type":55,"value":107},{"type":49,"tag":410,"props":1045,"children":1046},{},[1047,1049,1054],{"type":55,"value":1048},"If flow is not explicitly requested, default to ",{"type":49,"tag":92,"props":1050,"children":1052},{"className":1051},[],[1053],{"type":55,"value":427},{"type":55,"value":107},{"type":49,"tag":410,"props":1056,"children":1057},{},[1058],{"type":55,"value":1059},"If endpoint mode is unspecified, default to in-cluster persistent NIM reuse and\nautomatic NIM deploy\u002Frepair when unhealthy.",{"type":49,"tag":410,"props":1061,"children":1062},{},[1063,1065,1071],{"type":55,"value":1064},"If cache is missing, run ",{"type":49,"tag":92,"props":1066,"children":1068},{"className":1067},[],[1069],{"type":55,"value":1070},"setup_model_cache.yaml",{"type":55,"value":1072},", rerun pre-submit guard, and\ncontinue automatically on success.",{"type":49,"tag":410,"props":1074,"children":1075},{},[1076],{"type":55,"value":1077},"After any stage completes successfully, continue to the next stage immediately.\nDo not pause with \"Ready when you are\" or equivalent approval prompts.",{"type":49,"tag":912,"props":1079,"children":1081},{"id":1080},"triggers-that-should-pause-for-disambiguation",[1082],{"type":55,"value":1083},"Triggers that should pause for disambiguation",{"type":49,"tag":109,"props":1085,"children":1086},{},[1087,1108],{"type":49,"tag":113,"props":1088,"children":1089},{},[1090],{"type":49,"tag":117,"props":1091,"children":1092},{},[1093,1098,1103],{"type":49,"tag":121,"props":1094,"children":1095},{},[1096],{"type":55,"value":1097},"Missing input",{"type":49,"tag":121,"props":1099,"children":1100},{},[1101],{"type":55,"value":1102},"Why it matters",{"type":49,"tag":121,"props":1104,"children":1105},{},[1106],{"type":55,"value":1107},"Ask",{"type":49,"tag":137,"props":1109,"children":1110},{},[1111,1135,1153],{"type":49,"tag":117,"props":1112,"children":1113},{},[1114,1125,1130],{"type":49,"tag":144,"props":1115,"children":1116},{},[1117,1123],{"type":49,"tag":92,"props":1118,"children":1120},{"className":1119},[],[1121],{"type":55,"value":1122},"USER_INPUT_REQUIRED",{"type":55,"value":1124}," from preflight",{"type":49,"tag":144,"props":1126,"children":1127},{},[1128],{"type":55,"value":1129},"Required secret is missing",{"type":49,"tag":144,"props":1131,"children":1132},{},[1133],{"type":55,"value":1134},"Ask one concise unblock question for exactly the missing value(s)",{"type":49,"tag":117,"props":1136,"children":1137},{},[1138,1143,1148],{"type":49,"tag":144,"props":1139,"children":1140},{},[1141],{"type":55,"value":1142},"Storage backend prefix cannot be derived from the active dataset\u002Fupload root",{"type":49,"tag":144,"props":1144,"children":1145},{},[1146],{"type":55,"value":1147},"Wrong scheme causes runtime storage auth mismatch",{"type":49,"tag":144,"props":1149,"children":1150},{},[1151],{"type":55,"value":1152},"\"What is the backend-native root prefix for this run?\"",{"type":49,"tag":117,"props":1154,"children":1155},{},[1156,1161,1166],{"type":49,"tag":144,"props":1157,"children":1158},{},[1159],{"type":55,"value":1160},"No ONLINE GPU pool\u002Fplatform can be selected",{"type":49,"tag":144,"props":1162,"children":1163},{},[1164],{"type":55,"value":1165},"Workflow cannot schedule setup\u002Fworkers",{"type":49,"tag":144,"props":1167,"children":1168},{},[1169],{"type":55,"value":1170},"\"Which GPU pool\u002Fplatform should this run target?\"",{"type":49,"tag":912,"props":1172,"children":1174},{"id":1173},"when-not-to-disambiguate",[1175],{"type":55,"value":1176},"When NOT to disambiguate",{"type":49,"tag":1022,"props":1178,"children":1179},{},[1180,1185,1190,1195,1200,1205],{"type":49,"tag":410,"props":1181,"children":1182},{},[1183],{"type":55,"value":1184},"Do not ask for cookbook unless user explicitly asks to change scene profile.",{"type":49,"tag":410,"props":1186,"children":1187},{},[1188],{"type":55,"value":1189},"Do not offer external endpoints by default.",{"type":49,"tag":410,"props":1191,"children":1192},{},[1193],{"type":55,"value":1194},"Do not ask A\u002FB cache strategy questions; default is automatic cache setup.",{"type":49,"tag":410,"props":1196,"children":1197},{},[1198],{"type":55,"value":1199},"Do not ask to scale down existing NIMs; this is forbidden.",{"type":49,"tag":410,"props":1201,"children":1202},{},[1203],{"type":55,"value":1204},"Do not invent, scrape, or generate random videos when input is missing.",{"type":49,"tag":410,"props":1206,"children":1207},{},[1208],{"type":55,"value":1209},"Do not use non-VDA demo sources (for example Carline adaptation assets) unless\nthe user explicitly requests a different dataset.",{"type":49,"tag":64,"props":1211,"children":1213},{"id":1212},"step-0-select-flow-and-gather-inputs",[1214],{"type":55,"value":1215},"Step 0: Select Flow and Gather Inputs",{"type":49,"tag":912,"props":1217,"children":1219},{"id":1218},"input-video-policy-non-negotiable",[1220],{"type":55,"value":1221},"Input video policy (non-negotiable)",{"type":49,"tag":1022,"props":1223,"children":1224},{},[1225,1230,1235,1247,1252,1265],{"type":49,"tag":410,"props":1226,"children":1227},{},[1228],{"type":55,"value":1229},"Always preserve user-provided video inputs (dataset URL, local path, or upload\nfolder) as first-class and preferred.",{"type":49,"tag":410,"props":1231,"children":1232},{},[1233],{"type":55,"value":1234},"Never replace an explicit user video with demo assets or any other source.",{"type":49,"tag":410,"props":1236,"children":1237},{},[1238,1240,1245],{"type":55,"value":1239},"If no video input is provided, default to VDA demo assets via\n",{"type":49,"tag":92,"props":1241,"children":1243},{"className":1242},[],[1244],{"type":55,"value":615},{"type":55,"value":1246}," (HF dataset flow) without asking extra\nsource-selection questions.",{"type":49,"tag":410,"props":1248,"children":1249},{},[1250],{"type":55,"value":1251},"If the user explicitly mentions an input video or dataset, prefer and use that\ninput instead of demo assets.",{"type":49,"tag":410,"props":1253,"children":1254},{},[1255,1257,1263],{"type":55,"value":1256},"Use only VDA demo assets (",{"type":49,"tag":92,"props":1258,"children":1260},{"className":1259},[],[1261],{"type":55,"value":1262},"nvidia\u002Fvideo-data-augmentation-demo",{"type":55,"value":1264},") for the\ndefault demo path.",{"type":49,"tag":410,"props":1266,"children":1267},{},[1268],{"type":55,"value":1269},"Never propose arbitrary web clip downloads or placeholder videos\nunless the user explicitly requests that behavior.",{"type":49,"tag":58,"props":1271,"children":1272},{},[1273],{"type":55,"value":1274},"Collect only missing values:",{"type":49,"tag":406,"props":1276,"children":1277},{},[1278,1291,1328,1341],{"type":49,"tag":410,"props":1279,"children":1280},{},[1281,1283,1289],{"type":55,"value":1282},"Dataset source (prefer explicit user-provided ",{"type":49,"tag":92,"props":1284,"children":1286},{"className":1285},[],[1287],{"type":55,"value":1288},"dataset_url",{"type":55,"value":1290}," or local upload\nfolder; otherwise default to VDA demo assets and proceed).",{"type":49,"tag":410,"props":1292,"children":1293},{},[1294,1296,1301,1302,1307,1308,1313,1314,1319,1321,1326],{"type":55,"value":1295},"Flow (",{"type":49,"tag":92,"props":1297,"children":1299},{"className":1298},[],[1300],{"type":55,"value":420},{"type":55,"value":159},{"type":49,"tag":92,"props":1303,"children":1305},{"className":1304},[],[1306],{"type":55,"value":427},{"type":55,"value":159},{"type":49,"tag":92,"props":1309,"children":1311},{"className":1310},[],[1312],{"type":55,"value":434},{"type":55,"value":159},{"type":49,"tag":92,"props":1315,"children":1317},{"className":1316},[],[1318],{"type":55,"value":442},{"type":55,"value":1320},");\ndefault to ",{"type":49,"tag":92,"props":1322,"children":1324},{"className":1323},[],[1325],{"type":55,"value":427},{"type":55,"value":1327}," when unspecified.",{"type":49,"tag":410,"props":1329,"children":1330},{},[1331,1333,1339],{"type":55,"value":1332},"OSMO ",{"type":49,"tag":92,"props":1334,"children":1336},{"className":1335},[],[1337],{"type":55,"value":1338},"gpu_platform",{"type":55,"value":1340}," for all VDA resources (auto-select an ONLINE platform\nwhen unambiguous; ask only when no valid option exists).",{"type":49,"tag":410,"props":1342,"children":1343},{},[1344],{"type":55,"value":1345},"Endpoint mode (default in-cluster NIM reuse\u002Fdeploy unless explicitly\noverridden).",{"type":49,"tag":58,"props":1347,"children":1348},{},[1349,1351,1356,1358,1364,1366,1371,1372,1378],{"type":55,"value":1350},"Do not guess ",{"type":49,"tag":92,"props":1352,"children":1354},{"className":1353},[],[1355],{"type":55,"value":1338},{"type":55,"value":1357}," (for example ",{"type":49,"tag":92,"props":1359,"children":1361},{"className":1360},[],[1362],{"type":55,"value":1363},"microk8s",{"type":55,"value":1365},"). Use the exact current\nplatform label shown by ",{"type":49,"tag":92,"props":1367,"children":1369},{"className":1368},[],[1370],{"type":55,"value":327},{"type":55,"value":1357},{"type":49,"tag":92,"props":1373,"children":1375},{"className":1374},[],[1376],{"type":55,"value":1377},"gpu",{"type":55,"value":466},{"type":49,"tag":58,"props":1380,"children":1381},{},[1382],{"type":55,"value":1383},"Generate run stamp before each submit:",{"type":49,"tag":491,"props":1385,"children":1389},{"className":1386,"code":1387,"language":1388,"meta":496,"style":496},"language-bash shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","STAMP=$(cat \u002Fproc\u002Fsys\u002Fkernel\u002Frandom\u002Fuuid | cut -c1-8)\nRUN_ID=\"run-$STAMP\"\n","bash",[1390],{"type":49,"tag":92,"props":1391,"children":1392},{"__ignoreMap":496},[1393,1440],{"type":49,"tag":502,"props":1394,"children":1395},{"class":504,"line":505},[1396,1402,1408,1414,1420,1425,1430,1435],{"type":49,"tag":502,"props":1397,"children":1399},{"style":1398},"--shiki-light:#90A4AE;--shiki-default:#EEFFFF;--shiki-dark:#BABED8",[1400],{"type":55,"value":1401},"STAMP",{"type":49,"tag":502,"props":1403,"children":1405},{"style":1404},"--shiki-light:#39ADB5;--shiki-default:#89DDFF;--shiki-dark:#89DDFF",[1406],{"type":55,"value":1407},"=$(",{"type":49,"tag":502,"props":1409,"children":1411},{"style":1410},"--shiki-light:#E2931D;--shiki-default:#FFCB6B;--shiki-dark:#FFCB6B",[1412],{"type":55,"value":1413},"cat",{"type":49,"tag":502,"props":1415,"children":1417},{"style":1416},"--shiki-light:#91B859;--shiki-default:#C3E88D;--shiki-dark:#C3E88D",[1418],{"type":55,"value":1419}," \u002Fproc\u002Fsys\u002Fkernel\u002Frandom\u002Fuuid",{"type":49,"tag":502,"props":1421,"children":1422},{"style":1404},[1423],{"type":55,"value":1424}," |",{"type":49,"tag":502,"props":1426,"children":1427},{"style":1410},[1428],{"type":55,"value":1429}," cut",{"type":49,"tag":502,"props":1431,"children":1432},{"style":1416},[1433],{"type":55,"value":1434}," -c1-8",{"type":49,"tag":502,"props":1436,"children":1437},{"style":1404},[1438],{"type":55,"value":1439},")\n",{"type":49,"tag":502,"props":1441,"children":1442},{"class":504,"line":514},[1443,1448,1453,1458,1463,1468],{"type":49,"tag":502,"props":1444,"children":1445},{"style":1398},[1446],{"type":55,"value":1447},"RUN_ID",{"type":49,"tag":502,"props":1449,"children":1450},{"style":1404},[1451],{"type":55,"value":1452},"=",{"type":49,"tag":502,"props":1454,"children":1455},{"style":1404},[1456],{"type":55,"value":1457},"\"",{"type":49,"tag":502,"props":1459,"children":1460},{"style":1416},[1461],{"type":55,"value":1462},"run-",{"type":49,"tag":502,"props":1464,"children":1465},{"style":1398},[1466],{"type":55,"value":1467},"$STAMP",{"type":49,"tag":502,"props":1469,"children":1470},{"style":1404},[1471],{"type":55,"value":1472},"\"\n",{"type":49,"tag":64,"props":1474,"children":1476},{"id":1475},"execution-time-overview-required-before-run",[1477],{"type":55,"value":1478},"Execution Time Overview (required before run)",{"type":49,"tag":58,"props":1480,"children":1481},{},[1482,1484,1490],{"type":55,"value":1483},"Before running any mutating command (",{"type":49,"tag":92,"props":1485,"children":1487},{"className":1486},[],[1488],{"type":55,"value":1489},"osmo credential set",{"type":55,"value":1491},", NIM install\u002Frepair,\ncache workflow submit, or target VDA workflow submit), provide a short ETA\noverview to the user.",{"type":49,"tag":58,"props":1493,"children":1494},{},[1495],{"type":55,"value":1496},"Keep it concise (one short paragraph or 4-6 bullets) and include:",{"type":49,"tag":1022,"props":1498,"children":1499},{},[1500,1520,1525],{"type":49,"tag":410,"props":1501,"children":1502},{},[1503,1505,1511,1513,1518],{"type":55,"value":1504},"whether this looks like a ",{"type":49,"tag":1506,"props":1507,"children":1508},"strong",{},[1509],{"type":55,"value":1510},"cold start",{"type":55,"value":1512}," (NIM\u002Fcache missing) or ",{"type":49,"tag":1506,"props":1514,"children":1515},{},[1516],{"type":55,"value":1517},"warm start",{"type":55,"value":1519},"\n(NIM\u002Fcache already healthy),",{"type":49,"tag":410,"props":1521,"children":1522},{},[1523],{"type":55,"value":1524},"major phases with approximate durations,",{"type":49,"tag":410,"props":1526,"children":1527},{},[1528],{"type":55,"value":1529},"a total expected range for the selected workflow.",{"type":49,"tag":58,"props":1531,"children":1532},{},[1533],{"type":55,"value":1534},"Baseline ranges (from observed MicroK8s + OSMO runs):",{"type":49,"tag":109,"props":1536,"children":1537},{},[1538,1554],{"type":49,"tag":113,"props":1539,"children":1540},{},[1541],{"type":49,"tag":117,"props":1542,"children":1543},{},[1544,1549],{"type":49,"tag":121,"props":1545,"children":1546},{},[1547],{"type":55,"value":1548},"Phase",{"type":49,"tag":121,"props":1550,"children":1551},{},[1552],{"type":55,"value":1553},"Typical duration",{"type":49,"tag":137,"props":1555,"children":1556},{},[1557,1570,1583,1596,1609],{"type":49,"tag":117,"props":1558,"children":1559},{},[1560,1565],{"type":49,"tag":144,"props":1561,"children":1562},{},[1563],{"type":55,"value":1564},"Credentials + preflight",{"type":49,"tag":144,"props":1566,"children":1567},{},[1568],{"type":55,"value":1569},"~1-2 min",{"type":49,"tag":117,"props":1571,"children":1572},{},[1573,1578],{"type":49,"tag":144,"props":1574,"children":1575},{},[1576],{"type":55,"value":1577},"NIM deploy\u002Fdownload\u002Fwarmup (if needed)",{"type":49,"tag":144,"props":1579,"children":1580},{},[1581],{"type":55,"value":1582},"~10-15 min",{"type":49,"tag":117,"props":1584,"children":1585},{},[1586,1591],{"type":49,"tag":144,"props":1587,"children":1588},{},[1589],{"type":55,"value":1590},"Demo assets download\u002Fupload (if demo path)",{"type":49,"tag":144,"props":1592,"children":1593},{},[1594],{"type":55,"value":1595},"~1-3 min",{"type":49,"tag":117,"props":1597,"children":1598},{},[1599,1604],{"type":49,"tag":144,"props":1600,"children":1601},{},[1602],{"type":55,"value":1603},"Model cache population (if needed)",{"type":49,"tag":144,"props":1605,"children":1606},{},[1607],{"type":55,"value":1608},"~15-25 min",{"type":49,"tag":117,"props":1610,"children":1611},{},[1612,1617],{"type":49,"tag":144,"props":1613,"children":1614},{},[1615],{"type":55,"value":1616},"Workflow submit + queue\u002Fstart",{"type":49,"tag":144,"props":1618,"children":1619},{},[1620],{"type":55,"value":1595},{"type":49,"tag":58,"props":1622,"children":1623},{},[1624],{"type":55,"value":1625},"Workflow runtime ranges after submit:",{"type":49,"tag":109,"props":1627,"children":1628},{},[1629,1644],{"type":49,"tag":113,"props":1630,"children":1631},{},[1632],{"type":49,"tag":117,"props":1633,"children":1634},{},[1635,1639],{"type":49,"tag":121,"props":1636,"children":1637},{},[1638],{"type":55,"value":759},{"type":49,"tag":121,"props":1640,"children":1641},{},[1642],{"type":55,"value":1643},"Typical runtime",{"type":49,"tag":137,"props":1645,"children":1646},{},[1647,1663,1679,1695],{"type":49,"tag":117,"props":1648,"children":1649},{},[1650,1658],{"type":49,"tag":144,"props":1651,"children":1652},{},[1653],{"type":49,"tag":92,"props":1654,"children":1656},{"className":1655},[],[1657],{"type":55,"value":420},{"type":49,"tag":144,"props":1659,"children":1660},{},[1661],{"type":55,"value":1662},"~6-15 min",{"type":49,"tag":117,"props":1664,"children":1665},{},[1666,1674],{"type":49,"tag":144,"props":1667,"children":1668},{},[1669],{"type":49,"tag":92,"props":1670,"children":1672},{"className":1671},[],[1673],{"type":55,"value":427},{"type":49,"tag":144,"props":1675,"children":1676},{},[1677],{"type":55,"value":1678},"~20-35 min",{"type":49,"tag":117,"props":1680,"children":1681},{},[1682,1690],{"type":49,"tag":144,"props":1683,"children":1684},{},[1685],{"type":49,"tag":92,"props":1686,"children":1688},{"className":1687},[],[1689],{"type":55,"value":434},{"type":49,"tag":144,"props":1691,"children":1692},{},[1693],{"type":55,"value":1694},"~22-40 min",{"type":49,"tag":117,"props":1696,"children":1697},{},[1698,1706],{"type":49,"tag":144,"props":1699,"children":1700},{},[1701],{"type":49,"tag":92,"props":1702,"children":1704},{"className":1703},[],[1705],{"type":55,"value":442},{"type":49,"tag":144,"props":1707,"children":1708},{},[1709],{"type":55,"value":1710},"~25-45 min",{"type":49,"tag":58,"props":1712,"children":1713},{},[1714],{"type":55,"value":1715},"Cold-start end-to-end runs are commonly ~45-80 min; warm-start runs are usually\n~20-45 min depending on flow and video length.",{"type":49,"tag":64,"props":1717,"children":1719},{"id":1718},"common-preconditions-all-flows",[1720],{"type":55,"value":1721},"Common Preconditions (all flows)",{"type":49,"tag":406,"props":1723,"children":1724},{},[1725,2031,2072,2170,2275,2378],{"type":49,"tag":410,"props":1726,"children":1727},{},[1728,1733,1787,1791,1793,1844,1847,1849,1855,1857,1863,1864,1870,1871,1876,1877,1882,1883,1888,1889,1895,1897,1903,1905,1908,1910,1916,1918,1924,1926,1931,1932,1937,1939,1945,1947,1999,2002,2004,2009,2011,2014,2016,2022,2024,2029],{"type":49,"tag":1506,"props":1729,"children":1730},{},[1731],{"type":55,"value":1732},"Credential and control-plane preflight",{"type":49,"tag":491,"props":1734,"children":1736},{"className":1386,"code":1735,"language":1388,"meta":496,"style":496},"bash scripts\u002Fpreflight_credentials.sh --workflow assets\u002Fconfigs\u002Fosmo\u002F\u003Cmode>.yaml\n",[1737],{"type":49,"tag":92,"props":1738,"children":1739},{"__ignoreMap":496},[1740],{"type":49,"tag":502,"props":1741,"children":1742},{"class":504,"line":505},[1743,1747,1752,1757,1762,1767,1772,1777,1782],{"type":49,"tag":502,"props":1744,"children":1745},{"style":1410},[1746],{"type":55,"value":1388},{"type":49,"tag":502,"props":1748,"children":1749},{"style":1416},[1750],{"type":55,"value":1751}," scripts\u002Fpreflight_credentials.sh",{"type":49,"tag":502,"props":1753,"children":1754},{"style":1416},[1755],{"type":55,"value":1756}," --workflow",{"type":49,"tag":502,"props":1758,"children":1759},{"style":1416},[1760],{"type":55,"value":1761}," assets\u002Fconfigs\u002Fosmo\u002F",{"type":49,"tag":502,"props":1763,"children":1764},{"style":1404},[1765],{"type":55,"value":1766},"\u003C",{"type":49,"tag":502,"props":1768,"children":1769},{"style":1416},[1770],{"type":55,"value":1771},"mod",{"type":49,"tag":502,"props":1773,"children":1774},{"style":1398},[1775],{"type":55,"value":1776},"e",{"type":49,"tag":502,"props":1778,"children":1779},{"style":1404},[1780],{"type":55,"value":1781},">",{"type":49,"tag":502,"props":1783,"children":1784},{"style":1416},[1785],{"type":55,"value":1786},".yaml\n",{"type":49,"tag":1788,"props":1789,"children":1790},"br",{},[],{"type":55,"value":1792},"Restricted egress:",{"type":49,"tag":491,"props":1794,"children":1796},{"className":1386,"code":1795,"language":1388,"meta":496,"style":496},"bash scripts\u002Fpreflight_credentials.sh --no-probe --workflow assets\u002Fconfigs\u002Fosmo\u002F\u003Cmode>.yaml\n",[1797],{"type":49,"tag":92,"props":1798,"children":1799},{"__ignoreMap":496},[1800],{"type":49,"tag":502,"props":1801,"children":1802},{"class":504,"line":505},[1803,1807,1811,1816,1820,1824,1828,1832,1836,1840],{"type":49,"tag":502,"props":1804,"children":1805},{"style":1410},[1806],{"type":55,"value":1388},{"type":49,"tag":502,"props":1808,"children":1809},{"style":1416},[1810],{"type":55,"value":1751},{"type":49,"tag":502,"props":1812,"children":1813},{"style":1416},[1814],{"type":55,"value":1815}," --no-probe",{"type":49,"tag":502,"props":1817,"children":1818},{"style":1416},[1819],{"type":55,"value":1756},{"type":49,"tag":502,"props":1821,"children":1822},{"style":1416},[1823],{"type":55,"value":1761},{"type":49,"tag":502,"props":1825,"children":1826},{"style":1404},[1827],{"type":55,"value":1766},{"type":49,"tag":502,"props":1829,"children":1830},{"style":1416},[1831],{"type":55,"value":1771},{"type":49,"tag":502,"props":1833,"children":1834},{"style":1398},[1835],{"type":55,"value":1776},{"type":49,"tag":502,"props":1837,"children":1838},{"style":1404},[1839],{"type":55,"value":1781},{"type":49,"tag":502,"props":1841,"children":1842},{"style":1416},[1843],{"type":55,"value":1786},{"type":49,"tag":1788,"props":1845,"children":1846},{},[],{"type":55,"value":1848},"Preflight does not require a workload-local ",{"type":49,"tag":92,"props":1850,"children":1852},{"className":1851},[],[1853],{"type":55,"value":1854},".env",{"type":55,"value":1856},". Runtime interpolation is\ndriven by submit-time values (",{"type":49,"tag":92,"props":1858,"children":1860},{"className":1859},[],[1861],{"type":55,"value":1862},"dataset",{"type":55,"value":159},{"type":49,"tag":92,"props":1865,"children":1867},{"className":1866},[],[1868],{"type":55,"value":1869},"run_id",{"type":55,"value":159},{"type":49,"tag":92,"props":1872,"children":1874},{"className":1873},[],[1875],{"type":55,"value":1338},{"type":55,"value":159},{"type":49,"tag":92,"props":1878,"children":1880},{"className":1879},[],[1881],{"type":55,"value":22},{"type":55,"value":436},{"type":49,"tag":92,"props":1884,"children":1886},{"className":1885},[],[1887],{"type":55,"value":295},{"type":55,"value":159},{"type":49,"tag":92,"props":1890,"children":1892},{"className":1891},[],[1893],{"type":55,"value":1894},"skills_dir",{"type":55,"value":1896},") supplied in one ",{"type":49,"tag":92,"props":1898,"children":1900},{"className":1899},[],[1901],{"type":55,"value":1902},"--set-string",{"type":55,"value":1904}," list.",{"type":49,"tag":1788,"props":1906,"children":1907},{},[],{"type":55,"value":1909},"Passing ",{"type":49,"tag":92,"props":1911,"children":1913},{"className":1912},[],[1914],{"type":55,"value":1915},"--workflow",{"type":55,"value":1917}," validates pull access for the active workflow image refs\n(",{"type":49,"tag":92,"props":1919,"children":1921},{"className":1920},[],[1922],{"type":55,"value":1923},"workflow.groups[].tasks[].image",{"type":55,"value":1925},") using anonymous bearer access with\ncredential fallback when provided.\nIf replacement NGC\u002FHF secrets are provided in env, preflight refreshes\nexisting ",{"type":49,"tag":92,"props":1927,"children":1929},{"className":1928},[],[1930],{"type":55,"value":214},{"type":55,"value":371},{"type":49,"tag":92,"props":1933,"children":1935},{"className":1934},[],[1936],{"type":55,"value":260},{"type":55,"value":1938}," automatically when present. Use ",{"type":49,"tag":92,"props":1940,"children":1942},{"className":1941},[],[1943],{"type":55,"value":1944},"--refresh",{"type":55,"value":1946}," to force\noverwrite even when no new env secrets were supplied:",{"type":49,"tag":491,"props":1948,"children":1950},{"className":1386,"code":1949,"language":1388,"meta":496,"style":496},"bash scripts\u002Fpreflight_credentials.sh --workflow assets\u002Fconfigs\u002Fosmo\u002F\u003Cmode>.yaml --refresh\n",[1951],{"type":49,"tag":92,"props":1952,"children":1953},{"__ignoreMap":496},[1954],{"type":49,"tag":502,"props":1955,"children":1956},{"class":504,"line":505},[1957,1961,1965,1969,1973,1977,1981,1985,1989,1994],{"type":49,"tag":502,"props":1958,"children":1959},{"style":1410},[1960],{"type":55,"value":1388},{"type":49,"tag":502,"props":1962,"children":1963},{"style":1416},[1964],{"type":55,"value":1751},{"type":49,"tag":502,"props":1966,"children":1967},{"style":1416},[1968],{"type":55,"value":1756},{"type":49,"tag":502,"props":1970,"children":1971},{"style":1416},[1972],{"type":55,"value":1761},{"type":49,"tag":502,"props":1974,"children":1975},{"style":1404},[1976],{"type":55,"value":1766},{"type":49,"tag":502,"props":1978,"children":1979},{"style":1416},[1980],{"type":55,"value":1771},{"type":49,"tag":502,"props":1982,"children":1983},{"style":1398},[1984],{"type":55,"value":1776},{"type":49,"tag":502,"props":1986,"children":1987},{"style":1404},[1988],{"type":55,"value":1781},{"type":49,"tag":502,"props":1990,"children":1991},{"style":1416},[1992],{"type":55,"value":1993},".yaml",{"type":49,"tag":502,"props":1995,"children":1996},{"style":1416},[1997],{"type":55,"value":1998}," --refresh\n",{"type":49,"tag":1788,"props":2000,"children":2001},{},[],{"type":55,"value":2003},"If output contains ",{"type":49,"tag":92,"props":2005,"children":2007},{"className":2006},[],[2008],{"type":55,"value":97},{"type":55,"value":2010},", ask one concise unblock question\nand stop.",{"type":49,"tag":1788,"props":2012,"children":2013},{},[],{"type":55,"value":2015},"On workflow image ",{"type":49,"tag":92,"props":2017,"children":2019},{"className":2018},[],[2020],{"type":55,"value":2021},"401\u002F403",{"type":55,"value":2023},", report registry access failure after probe\nchecks on the listed image refs; do not claim a key family (for example\n",{"type":49,"tag":92,"props":2025,"children":2027},{"className":2026},[],[2028],{"type":55,"value":173},{"type":55,"value":2030},") is categorically unsupported.",{"type":49,"tag":410,"props":2032,"children":2033},{},[2034,2039,2042,2047,2049,2059,2062,2064,2070],{"type":49,"tag":1506,"props":2035,"children":2036},{},[2037],{"type":55,"value":2038},"Storage interpolation policy",{"type":49,"tag":1788,"props":2040,"children":2041},{},[],{"type":49,"tag":92,"props":2043,"children":2045},{"className":2044},[],[2046],{"type":55,"value":295},{"type":55,"value":2048}," must be derived from the actual dataset\u002Fupload backend for the\ncurrent run.",{"type":49,"tag":491,"props":2050,"children":2054},{"className":2051,"code":2053,"language":55,"meta":496},[2052],"language-text","dataset_url=azure:\u002F\u002Fstoriondevxah69\u002Fosmo-workflows\u002Fdatasets\u002Fvda-demo\nstorage_url=azure:\u002F\u002Fstoriondevxah69\u002Fosmo-workflows\ndataset=vda-demo\n",[2055],{"type":49,"tag":92,"props":2056,"children":2057},{"__ignoreMap":496},[2058],{"type":55,"value":2053},{"type":49,"tag":1788,"props":2060,"children":2061},{},[],{"type":55,"value":2063},"Never silently default to stale ",{"type":49,"tag":92,"props":2065,"children":2067},{"className":2066},[],[2068],{"type":55,"value":2069},"s3:\u002F\u002F",{"type":55,"value":2071}," values on non-S3 backends.",{"type":49,"tag":410,"props":2073,"children":2074},{},[2075,2080,2093,2139],{"type":49,"tag":1506,"props":2076,"children":2077},{},[2078],{"type":55,"value":2079},"Inference policy (non-negotiable)",{"type":49,"tag":1022,"props":2081,"children":2082},{},[2083,2088],{"type":49,"tag":410,"props":2084,"children":2085},{},[2086],{"type":55,"value":2087},"Reuse healthy in-cluster persistent NIM endpoints by default.",{"type":49,"tag":410,"props":2089,"children":2090},{},[2091],{"type":55,"value":2092},"If missing\u002Funhealthy, deploy automatically — this is a prerequisite, not a\nuser decision. Do NOT pause to ask; run the install with the VDA allow-list:",{"type":49,"tag":491,"props":2094,"children":2096},{"className":1386,"code":2095,"language":1388,"meta":496,"style":496},"export NIM_SERVICES=\"qwen3-vl qwen25-14b\"\nskills\u002Fphysical-ai-infrastructure-setup-and-resilient-scaling\u002Fcomponents\u002Finference-nim-operator\u002Fscripts\u002Finstall.sh\n",[2097],{"type":49,"tag":92,"props":2098,"children":2099},{"__ignoreMap":496},[2100,2131],{"type":49,"tag":502,"props":2101,"children":2102},{"class":504,"line":505},[2103,2109,2114,2118,2122,2127],{"type":49,"tag":502,"props":2104,"children":2106},{"style":2105},"--shiki-light:#9C3EDA;--shiki-default:#C792EA;--shiki-dark:#C792EA",[2107],{"type":55,"value":2108},"export",{"type":49,"tag":502,"props":2110,"children":2111},{"style":1398},[2112],{"type":55,"value":2113}," NIM_SERVICES",{"type":49,"tag":502,"props":2115,"children":2116},{"style":1404},[2117],{"type":55,"value":1452},{"type":49,"tag":502,"props":2119,"children":2120},{"style":1404},[2121],{"type":55,"value":1457},{"type":49,"tag":502,"props":2123,"children":2124},{"style":1416},[2125],{"type":55,"value":2126},"qwen3-vl qwen25-14b",{"type":49,"tag":502,"props":2128,"children":2129},{"style":1404},[2130],{"type":55,"value":1472},{"type":49,"tag":502,"props":2132,"children":2133},{"class":504,"line":514},[2134],{"type":49,"tag":502,"props":2135,"children":2136},{"style":1410},[2137],{"type":55,"value":2138},"skills\u002Fphysical-ai-infrastructure-setup-and-resilient-scaling\u002Fcomponents\u002Finference-nim-operator\u002Fscripts\u002Finstall.sh\n",{"type":49,"tag":1022,"props":2140,"children":2141},{},[2142,2155,2160,2165],{"type":49,"tag":410,"props":2143,"children":2144},{},[2145,2147,2153],{"type":55,"value":2146},"See ",{"type":49,"tag":92,"props":2148,"children":2150},{"className":2149},[],[2151],{"type":55,"value":2152},"references\u002Fnim\u002FREADME.md",{"type":55,"value":2154}," for full endpoint docs and health checks.",{"type":49,"tag":410,"props":2156,"children":2157},{},[2158],{"type":55,"value":2159},"External endpoints are opt-in only (explicit request or explicit URLs); only\nthen skip the in-cluster deploy.",{"type":49,"tag":410,"props":2161,"children":2162},{},[2163],{"type":55,"value":2164},"Never infer external mode from credential presence.",{"type":49,"tag":410,"props":2166,"children":2167},{},[2168],{"type":55,"value":2169},"Never scale down\u002Fdelete existing NIMs to free GPUs.",{"type":49,"tag":410,"props":2171,"children":2172},{},[2173,2178],{"type":49,"tag":1506,"props":2174,"children":2175},{},[2176],{"type":55,"value":2177},"Readiness guard",{"type":49,"tag":491,"props":2179,"children":2181},{"className":1386,"code":2180,"language":1388,"meta":496,"style":496},"osmo pool list --mode free\nosmo config show POD_TEMPLATE\npython3 scripts\u002Fpre_submit_guard.py --workflow assets\u002Fconfigs\u002Fosmo\u002F\u003Cmode>.yaml\n",[2182],{"type":49,"tag":92,"props":2183,"children":2184},{"__ignoreMap":496},[2185,2212,2234],{"type":49,"tag":502,"props":2186,"children":2187},{"class":504,"line":505},[2188,2192,2197,2202,2207],{"type":49,"tag":502,"props":2189,"children":2190},{"style":1410},[2191],{"type":55,"value":279},{"type":49,"tag":502,"props":2193,"children":2194},{"style":1416},[2195],{"type":55,"value":2196}," pool",{"type":49,"tag":502,"props":2198,"children":2199},{"style":1416},[2200],{"type":55,"value":2201}," list",{"type":49,"tag":502,"props":2203,"children":2204},{"style":1416},[2205],{"type":55,"value":2206}," --mode",{"type":49,"tag":502,"props":2208,"children":2209},{"style":1416},[2210],{"type":55,"value":2211}," free\n",{"type":49,"tag":502,"props":2213,"children":2214},{"class":504,"line":514},[2215,2219,2224,2229],{"type":49,"tag":502,"props":2216,"children":2217},{"style":1410},[2218],{"type":55,"value":279},{"type":49,"tag":502,"props":2220,"children":2221},{"style":1416},[2222],{"type":55,"value":2223}," config",{"type":49,"tag":502,"props":2225,"children":2226},{"style":1416},[2227],{"type":55,"value":2228}," show",{"type":49,"tag":502,"props":2230,"children":2231},{"style":1416},[2232],{"type":55,"value":2233}," POD_TEMPLATE\n",{"type":49,"tag":502,"props":2235,"children":2236},{"class":504,"line":523},[2237,2242,2247,2251,2255,2259,2263,2267,2271],{"type":49,"tag":502,"props":2238,"children":2239},{"style":1410},[2240],{"type":55,"value":2241},"python3",{"type":49,"tag":502,"props":2243,"children":2244},{"style":1416},[2245],{"type":55,"value":2246}," scripts\u002Fpre_submit_guard.py",{"type":49,"tag":502,"props":2248,"children":2249},{"style":1416},[2250],{"type":55,"value":1756},{"type":49,"tag":502,"props":2252,"children":2253},{"style":1416},[2254],{"type":55,"value":1761},{"type":49,"tag":502,"props":2256,"children":2257},{"style":1404},[2258],{"type":55,"value":1766},{"type":49,"tag":502,"props":2260,"children":2261},{"style":1416},[2262],{"type":55,"value":1771},{"type":49,"tag":502,"props":2264,"children":2265},{"style":1398},[2266],{"type":55,"value":1776},{"type":49,"tag":502,"props":2268,"children":2269},{"style":1404},[2270],{"type":55,"value":1781},{"type":49,"tag":502,"props":2272,"children":2273},{"style":1416},[2274],{"type":55,"value":1786},{"type":49,"tag":410,"props":2276,"children":2277},{},[2278,2283,2286,2288,2294,2296,2366,2369,2371,2376],{"type":49,"tag":1506,"props":2279,"children":2280},{},[2281],{"type":55,"value":2282},"Cache auto-remediation",{"type":49,"tag":1788,"props":2284,"children":2285},{},[],{"type":55,"value":2287},"If ",{"type":49,"tag":92,"props":2289,"children":2291},{"className":2290},[],[2292],{"type":55,"value":2293},"pre_submit_guard.py",{"type":55,"value":2295}," reports cache failure, default action is to run:",{"type":49,"tag":491,"props":2297,"children":2299},{"className":1386,"code":2298,"language":1388,"meta":496,"style":496},"osmo workflow submit assets\u002Fconfigs\u002Fosmo\u002Fsetup_model_cache.yaml \\\n  --set-string storage_url=\u003Cbackend-prefix> path=data\n",[2300],{"type":49,"tag":92,"props":2301,"children":2302},{"__ignoreMap":496},[2303,2330],{"type":49,"tag":502,"props":2304,"children":2305},{"class":504,"line":505},[2306,2310,2315,2320,2325],{"type":49,"tag":502,"props":2307,"children":2308},{"style":1410},[2309],{"type":55,"value":279},{"type":49,"tag":502,"props":2311,"children":2312},{"style":1416},[2313],{"type":55,"value":2314}," workflow",{"type":49,"tag":502,"props":2316,"children":2317},{"style":1416},[2318],{"type":55,"value":2319}," submit",{"type":49,"tag":502,"props":2321,"children":2322},{"style":1416},[2323],{"type":55,"value":2324}," assets\u002Fconfigs\u002Fosmo\u002Fsetup_model_cache.yaml",{"type":49,"tag":502,"props":2326,"children":2327},{"style":1398},[2328],{"type":55,"value":2329}," \\\n",{"type":49,"tag":502,"props":2331,"children":2332},{"class":504,"line":514},[2333,2338,2343,2347,2352,2357,2361],{"type":49,"tag":502,"props":2334,"children":2335},{"style":1416},[2336],{"type":55,"value":2337},"  --set-string",{"type":49,"tag":502,"props":2339,"children":2340},{"style":1416},[2341],{"type":55,"value":2342}," storage_url=",{"type":49,"tag":502,"props":2344,"children":2345},{"style":1404},[2346],{"type":55,"value":1766},{"type":49,"tag":502,"props":2348,"children":2349},{"style":1416},[2350],{"type":55,"value":2351},"backend-prefi",{"type":49,"tag":502,"props":2353,"children":2354},{"style":1398},[2355],{"type":55,"value":2356},"x",{"type":49,"tag":502,"props":2358,"children":2359},{"style":1404},[2360],{"type":55,"value":1781},{"type":49,"tag":502,"props":2362,"children":2363},{"style":1416},[2364],{"type":55,"value":2365}," path=data\n",{"type":49,"tag":1788,"props":2367,"children":2368},{},[],{"type":55,"value":2370},"Then rerun ",{"type":49,"tag":92,"props":2372,"children":2374},{"className":2373},[],[2375],{"type":55,"value":2293},{"type":55,"value":2377}," and submit the target VDA flow only after it\npasses. Ask user only when backend\u002Fprefix is ambiguous or cache setup fails.",{"type":49,"tag":410,"props":2379,"children":2380},{},[2381,2386,2389,2391,2396,2398,2404],{"type":49,"tag":1506,"props":2382,"children":2383},{},[2384],{"type":55,"value":2385},"Scheduling policy",{"type":49,"tag":1788,"props":2387,"children":2388},{},[],{"type":55,"value":2390},"VDA templates schedule setup and workers on ",{"type":49,"tag":92,"props":2392,"children":2394},{"className":2393},[],[2395],{"type":55,"value":1338},{"type":55,"value":2397}," (no ",{"type":49,"tag":92,"props":2399,"children":2401},{"className":2400},[],[2402],{"type":55,"value":2403},"system",{"type":55,"value":2405}," pool\ndependency for user workloads).",{"type":49,"tag":64,"props":2407,"children":2409},{"id":2408},"submit-all-flows",[2410],{"type":55,"value":2411},"Submit (all flows)",{"type":49,"tag":58,"props":2413,"children":2414},{},[2415],{"type":55,"value":2416},"Every flow uses the same submit shape; only the workflow YAML changes. Choose the\nYAML for the requested flow, then run the command below. Full per-flow walkthroughs\n(stage matrix and flow details) live in the linked references.",{"type":49,"tag":109,"props":2418,"children":2419},{},[2420,2440],{"type":49,"tag":113,"props":2421,"children":2422},{},[2423],{"type":49,"tag":117,"props":2424,"children":2425},{},[2426,2430,2435],{"type":49,"tag":121,"props":2427,"children":2428},{},[2429],{"type":55,"value":759},{"type":49,"tag":121,"props":2431,"children":2432},{},[2433],{"type":55,"value":2434},"Workflow YAML",{"type":49,"tag":121,"props":2436,"children":2437},{},[2438],{"type":55,"value":2439},"Walkthrough",{"type":49,"tag":137,"props":2441,"children":2442},{},[2443,2468,2493,2518],{"type":49,"tag":117,"props":2444,"children":2445},{},[2446,2451,2459],{"type":49,"tag":144,"props":2447,"children":2448},{},[2449],{"type":55,"value":2450},"Augmentation + auto-labeling",{"type":49,"tag":144,"props":2452,"children":2453},{},[2454],{"type":49,"tag":92,"props":2455,"children":2457},{"className":2456},[],[2458],{"type":55,"value":797},{"type":49,"tag":144,"props":2460,"children":2461},{},[2462],{"type":49,"tag":92,"props":2463,"children":2465},{"className":2464},[],[2466],{"type":55,"value":2467},"references\u002Fflows\u002Faugmentation_and_al.md",{"type":49,"tag":117,"props":2469,"children":2470},{},[2471,2476,2484],{"type":49,"tag":144,"props":2472,"children":2473},{},[2474],{"type":55,"value":2475},"Auto-labeling only",{"type":49,"tag":144,"props":2477,"children":2478},{},[2479],{"type":49,"tag":92,"props":2480,"children":2482},{"className":2481},[],[2483],{"type":55,"value":827},{"type":49,"tag":144,"props":2485,"children":2486},{},[2487],{"type":49,"tag":92,"props":2488,"children":2490},{"className":2489},[],[2491],{"type":55,"value":2492},"references\u002Fflows\u002Fauto_labeling.md",{"type":49,"tag":117,"props":2494,"children":2495},{},[2496,2501,2509],{"type":49,"tag":144,"props":2497,"children":2498},{},[2499],{"type":55,"value":2500},"E2E (parallel)",{"type":49,"tag":144,"props":2502,"children":2503},{},[2504],{"type":49,"tag":92,"props":2505,"children":2507},{"className":2506},[],[2508],{"type":55,"value":857},{"type":49,"tag":144,"props":2510,"children":2511},{},[2512],{"type":49,"tag":92,"props":2513,"children":2515},{"className":2514},[],[2516],{"type":55,"value":2517},"references\u002Fflows\u002Fe2e.md",{"type":49,"tag":117,"props":2519,"children":2520},{},[2521,2526,2534],{"type":49,"tag":144,"props":2522,"children":2523},{},[2524],{"type":55,"value":2525},"E2E (super-resolution gated)",{"type":49,"tag":144,"props":2527,"children":2528},{},[2529],{"type":49,"tag":92,"props":2530,"children":2532},{"className":2531},[],[2533],{"type":55,"value":887},{"type":49,"tag":144,"props":2535,"children":2536},{},[2537],{"type":49,"tag":92,"props":2538,"children":2540},{"className":2539},[],[2541],{"type":55,"value":2542},"references\u002Fflows\u002Fe2e_super_resolution.md",{"type":49,"tag":491,"props":2544,"children":2546},{"className":1386,"code":2545,"language":1388,"meta":496,"style":496},"SKILLS_DIR=\"$(cd \"$(git rev-parse --show-toplevel)\u002Fskills\u002Fphysical-ai-video-data-augmentation\" && pwd)\"\nSTAMP=$(cat \u002Fproc\u002Fsys\u002Fkernel\u002Frandom\u002Fuuid | cut -c1-8)\nosmo workflow submit assets\u002Fconfigs\u002Fosmo\u002F\u003Cflow>.yaml \\\n  --pool \u003Cpool> \\\n  --set-string \\\n    dataset=\u003Cdataset> \\\n    run_id=run-$STAMP \\\n    storage_url=\u003Cbackend-prefix> \\\n    gpu_platform=\u003Cgpu-platform> \\\n    video=\u003Cvideo-stem> \\\n    cosmos_model_cache_url=\u003Cbackend-prefix>\u002Fdata\u002Fmodels\u002Fcosmos_transfer \\\n    auto_labeling_model_cache_url=\u003Cbackend-prefix>\u002Fdata\u002Fmodels\u002Fauto_labeling \\\n    skills_dir=\"$SKILLS_DIR\"\n",[2547],{"type":49,"tag":92,"props":2548,"children":2549},{"__ignoreMap":496},[2550,2617,2652,2697,2729,2741,2772,2786,2815,2846,2876,2910,2944],{"type":49,"tag":502,"props":2551,"children":2552},{"class":504,"line":505},[2553,2558,2562,2567,2573,2578,2583,2588,2593,2598,2602,2607,2612],{"type":49,"tag":502,"props":2554,"children":2555},{"style":1398},[2556],{"type":55,"value":2557},"SKILLS_DIR",{"type":49,"tag":502,"props":2559,"children":2560},{"style":1404},[2561],{"type":55,"value":1452},{"type":49,"tag":502,"props":2563,"children":2564},{"style":1404},[2565],{"type":55,"value":2566},"\"$(",{"type":49,"tag":502,"props":2568,"children":2570},{"style":2569},"--shiki-light:#6182B8;--shiki-default:#82AAFF;--shiki-dark:#82AAFF",[2571],{"type":55,"value":2572},"cd",{"type":49,"tag":502,"props":2574,"children":2575},{"style":1404},[2576],{"type":55,"value":2577}," \"$(",{"type":49,"tag":502,"props":2579,"children":2580},{"style":1410},[2581],{"type":55,"value":2582},"git",{"type":49,"tag":502,"props":2584,"children":2585},{"style":1416},[2586],{"type":55,"value":2587}," rev-parse --show-toplevel",{"type":49,"tag":502,"props":2589,"children":2590},{"style":1404},[2591],{"type":55,"value":2592},")",{"type":49,"tag":502,"props":2594,"children":2595},{"style":1416},[2596],{"type":55,"value":2597},"\u002Fskills\u002Fphysical-ai-video-data-augmentation",{"type":49,"tag":502,"props":2599,"children":2600},{"style":1404},[2601],{"type":55,"value":1457},{"type":49,"tag":502,"props":2603,"children":2604},{"style":1404},[2605],{"type":55,"value":2606}," &&",{"type":49,"tag":502,"props":2608,"children":2609},{"style":2569},[2610],{"type":55,"value":2611}," pwd",{"type":49,"tag":502,"props":2613,"children":2614},{"style":1404},[2615],{"type":55,"value":2616},")\"\n",{"type":49,"tag":502,"props":2618,"children":2619},{"class":504,"line":514},[2620,2624,2628,2632,2636,2640,2644,2648],{"type":49,"tag":502,"props":2621,"children":2622},{"style":1398},[2623],{"type":55,"value":1401},{"type":49,"tag":502,"props":2625,"children":2626},{"style":1404},[2627],{"type":55,"value":1407},{"type":49,"tag":502,"props":2629,"children":2630},{"style":1410},[2631],{"type":55,"value":1413},{"type":49,"tag":502,"props":2633,"children":2634},{"style":1416},[2635],{"type":55,"value":1419},{"type":49,"tag":502,"props":2637,"children":2638},{"style":1404},[2639],{"type":55,"value":1424},{"type":49,"tag":502,"props":2641,"children":2642},{"style":1410},[2643],{"type":55,"value":1429},{"type":49,"tag":502,"props":2645,"children":2646},{"style":1416},[2647],{"type":55,"value":1434},{"type":49,"tag":502,"props":2649,"children":2650},{"style":1404},[2651],{"type":55,"value":1439},{"type":49,"tag":502,"props":2653,"children":2654},{"class":504,"line":523},[2655,2659,2663,2667,2671,2675,2680,2685,2689,2693],{"type":49,"tag":502,"props":2656,"children":2657},{"style":1410},[2658],{"type":55,"value":279},{"type":49,"tag":502,"props":2660,"children":2661},{"style":1416},[2662],{"type":55,"value":2314},{"type":49,"tag":502,"props":2664,"children":2665},{"style":1416},[2666],{"type":55,"value":2319},{"type":49,"tag":502,"props":2668,"children":2669},{"style":1416},[2670],{"type":55,"value":1761},{"type":49,"tag":502,"props":2672,"children":2673},{"style":1404},[2674],{"type":55,"value":1766},{"type":49,"tag":502,"props":2676,"children":2677},{"style":1416},[2678],{"type":55,"value":2679},"flo",{"type":49,"tag":502,"props":2681,"children":2682},{"style":1398},[2683],{"type":55,"value":2684},"w",{"type":49,"tag":502,"props":2686,"children":2687},{"style":1404},[2688],{"type":55,"value":1781},{"type":49,"tag":502,"props":2690,"children":2691},{"style":1416},[2692],{"type":55,"value":1993},{"type":49,"tag":502,"props":2694,"children":2695},{"style":1398},[2696],{"type":55,"value":2329},{"type":49,"tag":502,"props":2698,"children":2700},{"class":504,"line":2699},4,[2701,2706,2711,2716,2721,2725],{"type":49,"tag":502,"props":2702,"children":2703},{"style":1416},[2704],{"type":55,"value":2705},"  --pool",{"type":49,"tag":502,"props":2707,"children":2708},{"style":1404},[2709],{"type":55,"value":2710}," \u003C",{"type":49,"tag":502,"props":2712,"children":2713},{"style":1416},[2714],{"type":55,"value":2715},"poo",{"type":49,"tag":502,"props":2717,"children":2718},{"style":1398},[2719],{"type":55,"value":2720},"l",{"type":49,"tag":502,"props":2722,"children":2723},{"style":1404},[2724],{"type":55,"value":1781},{"type":49,"tag":502,"props":2726,"children":2727},{"style":1398},[2728],{"type":55,"value":2329},{"type":49,"tag":502,"props":2730,"children":2732},{"class":504,"line":2731},5,[2733,2737],{"type":49,"tag":502,"props":2734,"children":2735},{"style":1416},[2736],{"type":55,"value":2337},{"type":49,"tag":502,"props":2738,"children":2739},{"style":1398},[2740],{"type":55,"value":2329},{"type":49,"tag":502,"props":2742,"children":2744},{"class":504,"line":2743},6,[2745,2750,2754,2759,2764,2768],{"type":49,"tag":502,"props":2746,"children":2747},{"style":1416},[2748],{"type":55,"value":2749},"    dataset=",{"type":49,"tag":502,"props":2751,"children":2752},{"style":1404},[2753],{"type":55,"value":1766},{"type":49,"tag":502,"props":2755,"children":2756},{"style":1416},[2757],{"type":55,"value":2758},"datase",{"type":49,"tag":502,"props":2760,"children":2761},{"style":1398},[2762],{"type":55,"value":2763},"t",{"type":49,"tag":502,"props":2765,"children":2766},{"style":1404},[2767],{"type":55,"value":1781},{"type":49,"tag":502,"props":2769,"children":2770},{"style":1398},[2771],{"type":55,"value":2329},{"type":49,"tag":502,"props":2773,"children":2775},{"class":504,"line":2774},7,[2776,2781],{"type":49,"tag":502,"props":2777,"children":2778},{"style":1416},[2779],{"type":55,"value":2780},"    run_id=run-",{"type":49,"tag":502,"props":2782,"children":2783},{"style":1398},[2784],{"type":55,"value":2785},"$STAMP \\\n",{"type":49,"tag":502,"props":2787,"children":2789},{"class":504,"line":2788},8,[2790,2795,2799,2803,2807,2811],{"type":49,"tag":502,"props":2791,"children":2792},{"style":1416},[2793],{"type":55,"value":2794},"    storage_url=",{"type":49,"tag":502,"props":2796,"children":2797},{"style":1404},[2798],{"type":55,"value":1766},{"type":49,"tag":502,"props":2800,"children":2801},{"style":1416},[2802],{"type":55,"value":2351},{"type":49,"tag":502,"props":2804,"children":2805},{"style":1398},[2806],{"type":55,"value":2356},{"type":49,"tag":502,"props":2808,"children":2809},{"style":1404},[2810],{"type":55,"value":1781},{"type":49,"tag":502,"props":2812,"children":2813},{"style":1398},[2814],{"type":55,"value":2329},{"type":49,"tag":502,"props":2816,"children":2818},{"class":504,"line":2817},9,[2819,2824,2828,2833,2838,2842],{"type":49,"tag":502,"props":2820,"children":2821},{"style":1416},[2822],{"type":55,"value":2823},"    gpu_platform=",{"type":49,"tag":502,"props":2825,"children":2826},{"style":1404},[2827],{"type":55,"value":1766},{"type":49,"tag":502,"props":2829,"children":2830},{"style":1416},[2831],{"type":55,"value":2832},"gpu-platfor",{"type":49,"tag":502,"props":2834,"children":2835},{"style":1398},[2836],{"type":55,"value":2837},"m",{"type":49,"tag":502,"props":2839,"children":2840},{"style":1404},[2841],{"type":55,"value":1781},{"type":49,"tag":502,"props":2843,"children":2844},{"style":1398},[2845],{"type":55,"value":2329},{"type":49,"tag":502,"props":2847,"children":2849},{"class":504,"line":2848},10,[2850,2855,2859,2864,2868,2872],{"type":49,"tag":502,"props":2851,"children":2852},{"style":1416},[2853],{"type":55,"value":2854},"    video=",{"type":49,"tag":502,"props":2856,"children":2857},{"style":1404},[2858],{"type":55,"value":1766},{"type":49,"tag":502,"props":2860,"children":2861},{"style":1416},[2862],{"type":55,"value":2863},"video-ste",{"type":49,"tag":502,"props":2865,"children":2866},{"style":1398},[2867],{"type":55,"value":2837},{"type":49,"tag":502,"props":2869,"children":2870},{"style":1404},[2871],{"type":55,"value":1781},{"type":49,"tag":502,"props":2873,"children":2874},{"style":1398},[2875],{"type":55,"value":2329},{"type":49,"tag":502,"props":2877,"children":2879},{"class":504,"line":2878},11,[2880,2885,2889,2893,2897,2901,2906],{"type":49,"tag":502,"props":2881,"children":2882},{"style":1416},[2883],{"type":55,"value":2884},"    cosmos_model_cache_url=",{"type":49,"tag":502,"props":2886,"children":2887},{"style":1404},[2888],{"type":55,"value":1766},{"type":49,"tag":502,"props":2890,"children":2891},{"style":1416},[2892],{"type":55,"value":2351},{"type":49,"tag":502,"props":2894,"children":2895},{"style":1398},[2896],{"type":55,"value":2356},{"type":49,"tag":502,"props":2898,"children":2899},{"style":1404},[2900],{"type":55,"value":1781},{"type":49,"tag":502,"props":2902,"children":2903},{"style":1416},[2904],{"type":55,"value":2905},"\u002Fdata\u002Fmodels\u002Fcosmos_transfer",{"type":49,"tag":502,"props":2907,"children":2908},{"style":1398},[2909],{"type":55,"value":2329},{"type":49,"tag":502,"props":2911,"children":2913},{"class":504,"line":2912},12,[2914,2919,2923,2927,2931,2935,2940],{"type":49,"tag":502,"props":2915,"children":2916},{"style":1416},[2917],{"type":55,"value":2918},"    auto_labeling_model_cache_url=",{"type":49,"tag":502,"props":2920,"children":2921},{"style":1404},[2922],{"type":55,"value":1766},{"type":49,"tag":502,"props":2924,"children":2925},{"style":1416},[2926],{"type":55,"value":2351},{"type":49,"tag":502,"props":2928,"children":2929},{"style":1398},[2930],{"type":55,"value":2356},{"type":49,"tag":502,"props":2932,"children":2933},{"style":1404},[2934],{"type":55,"value":1781},{"type":49,"tag":502,"props":2936,"children":2937},{"style":1416},[2938],{"type":55,"value":2939},"\u002Fdata\u002Fmodels\u002Fauto_labeling",{"type":49,"tag":502,"props":2941,"children":2942},{"style":1398},[2943],{"type":55,"value":2329},{"type":49,"tag":502,"props":2945,"children":2947},{"class":504,"line":2946},13,[2948,2953,2957,2962],{"type":49,"tag":502,"props":2949,"children":2950},{"style":1416},[2951],{"type":55,"value":2952},"    skills_dir=",{"type":49,"tag":502,"props":2954,"children":2955},{"style":1404},[2956],{"type":55,"value":1457},{"type":49,"tag":502,"props":2958,"children":2959},{"style":1398},[2960],{"type":55,"value":2961},"$SKILLS_DIR",{"type":49,"tag":502,"props":2963,"children":2964},{"style":1404},[2965],{"type":55,"value":1472},{"type":49,"tag":58,"props":2967,"children":2968},{},[2969],{"type":55,"value":2970},"Compatibility note:",{"type":49,"tag":1022,"props":2972,"children":2973},{},[2974,2986,3005,3024,3037],{"type":49,"tag":410,"props":2975,"children":2976},{},[2977,2979,2984],{"type":55,"value":2978},"Use exactly one ",{"type":49,"tag":92,"props":2980,"children":2982},{"className":2981},[],[2983],{"type":55,"value":1902},{"type":55,"value":2985}," flag and pass all the key\u002Fvalue pairs after it.",{"type":49,"tag":410,"props":2987,"children":2988},{},[2989,2991,2997,2998,3003],{"type":55,"value":2990},"Do not repeat ",{"type":49,"tag":92,"props":2992,"children":2994},{"className":2993},[],[2995],{"type":55,"value":2996},"--set",{"type":55,"value":183},{"type":49,"tag":92,"props":2999,"children":3001},{"className":3000},[],[3002],{"type":55,"value":1902},{"type":55,"value":3004}," flags in the same command; some OSMO builds\nonly honor the last occurrence.",{"type":49,"tag":410,"props":3006,"children":3007},{},[3008,3010,3015,3017,3022],{"type":55,"value":3009},"Do not mix ",{"type":49,"tag":92,"props":3011,"children":3013},{"className":3012},[],[3014],{"type":55,"value":2996},{"type":55,"value":3016}," and ",{"type":49,"tag":92,"props":3018,"children":3020},{"className":3019},[],[3021],{"type":55,"value":1902},{"type":55,"value":3023}," in one submit command.",{"type":49,"tag":410,"props":3025,"children":3026},{},[3027,3029,3035],{"type":55,"value":3028},"Pass explicit ",{"type":49,"tag":92,"props":3030,"children":3032},{"className":3031},[],[3033],{"type":55,"value":3034},"*_model_cache_url",{"type":55,"value":3036}," values to avoid nested-template interpolation\ndifferences across OSMO environments.",{"type":49,"tag":410,"props":3038,"children":3039},{},[3040],{"type":55,"value":3041},"Do not brute-force permutations of flags. Use this shape directly.",{"type":49,"tag":58,"props":3043,"children":3044},{},[3045,3047,3052],{"type":55,"value":3046},"Common optional overrides (append key\u002Fvalue pairs to the same ",{"type":49,"tag":92,"props":3048,"children":3050},{"className":3049},[],[3051],{"type":55,"value":1902},{"type":55,"value":3053}," list):",{"type":49,"tag":491,"props":3055,"children":3057},{"className":1386,"code":3056,"language":1388,"meta":496,"style":496},"cookbook=\u003Cscene_profile> \\\nvlm_url=\u003Copenai_base_url> \\\nllm_url=\u003Copenai_base_url> \\\ncosmos_model_cache_url=\u003Curl> \\\nauto_labeling_model_cache_url=\u003Curl>\n",[3058],{"type":49,"tag":92,"props":3059,"children":3060},{"__ignoreMap":496},[3061,3087,3095,3103,3111],{"type":49,"tag":502,"props":3062,"children":3063},{"class":504,"line":505},[3064,3069,3074,3079,3083],{"type":49,"tag":502,"props":3065,"children":3066},{"style":1398},[3067],{"type":55,"value":3068},"cookbook",{"type":49,"tag":502,"props":3070,"children":3071},{"style":1404},[3072],{"type":55,"value":3073},"=\u003C",{"type":49,"tag":502,"props":3075,"children":3076},{"style":1416},[3077],{"type":55,"value":3078},"scene_profile",{"type":49,"tag":502,"props":3080,"children":3081},{"style":1404},[3082],{"type":55,"value":1781},{"type":49,"tag":502,"props":3084,"children":3085},{"style":1410},[3086],{"type":55,"value":2329},{"type":49,"tag":502,"props":3088,"children":3089},{"class":504,"line":514},[3090],{"type":49,"tag":502,"props":3091,"children":3092},{"style":1398},[3093],{"type":55,"value":3094},"vlm_url=\u003Copenai_base_url> \\\n",{"type":49,"tag":502,"props":3096,"children":3097},{"class":504,"line":523},[3098],{"type":49,"tag":502,"props":3099,"children":3100},{"style":1398},[3101],{"type":55,"value":3102},"llm_url=\u003Copenai_base_url> \\\n",{"type":49,"tag":502,"props":3104,"children":3105},{"class":504,"line":2699},[3106],{"type":49,"tag":502,"props":3107,"children":3108},{"style":1398},[3109],{"type":55,"value":3110},"cosmos_model_cache_url=\u003Curl> \\\n",{"type":49,"tag":502,"props":3112,"children":3113},{"class":504,"line":2731},[3114],{"type":49,"tag":502,"props":3115,"children":3116},{"style":1398},[3117],{"type":55,"value":3118},"auto_labeling_model_cache_url=\u003Curl>\n",{"type":49,"tag":58,"props":3120,"children":3121},{},[3122,3124,3130],{"type":55,"value":3123},"The auto-labeling-only flow has no augmentation stage, so it omits\n",{"type":49,"tag":92,"props":3125,"children":3127},{"className":3126},[],[3128],{"type":55,"value":3129},"cosmos_model_cache_url",{"type":55,"value":3131}," at runtime; passing it is harmless and keeps one submit\nshape across flows.",{"type":49,"tag":64,"props":3133,"children":3135},{"id":3134},"osmo-monitoring",[3136],{"type":55,"value":3137},"OSMO Monitoring",{"type":49,"tag":491,"props":3139,"children":3141},{"className":1386,"code":3140,"language":1388,"meta":496,"style":496},"# Workflow status + task states\nosmo workflow query \u003Cworkflow_id> --format-type json \\\n  | jq '{status, tasks: [.groups[].tasks[] | {name, status, exit_code}]}'\n\n# Logs for a specific task\nosmo workflow logs \u003Cworkflow_id> --task \u003Ctask_name> -n 200\n\n# Output retrieval\nosmo data list --no-pager \u003Coutput_url>\nosmo data download \u003Coutput_url> \u003Clocal_dir>\u002F\n",[3142],{"type":49,"tag":92,"props":3143,"children":3144},{"__ignoreMap":496},[3145,3154,3202,3230,3239,3247,3312,3319,3327,3366],{"type":49,"tag":502,"props":3146,"children":3147},{"class":504,"line":505},[3148],{"type":49,"tag":502,"props":3149,"children":3151},{"style":3150},"--shiki-light:#90A4AE;--shiki-light-font-style:italic;--shiki-default:#546E7A;--shiki-default-font-style:italic;--shiki-dark:#676E95;--shiki-dark-font-style:italic",[3152],{"type":55,"value":3153},"# Workflow status + task states\n",{"type":49,"tag":502,"props":3155,"children":3156},{"class":504,"line":514},[3157,3161,3165,3170,3174,3179,3184,3188,3193,3198],{"type":49,"tag":502,"props":3158,"children":3159},{"style":1410},[3160],{"type":55,"value":279},{"type":49,"tag":502,"props":3162,"children":3163},{"style":1416},[3164],{"type":55,"value":2314},{"type":49,"tag":502,"props":3166,"children":3167},{"style":1416},[3168],{"type":55,"value":3169}," query",{"type":49,"tag":502,"props":3171,"children":3172},{"style":1404},[3173],{"type":55,"value":2710},{"type":49,"tag":502,"props":3175,"children":3176},{"style":1416},[3177],{"type":55,"value":3178},"workflow_i",{"type":49,"tag":502,"props":3180,"children":3181},{"style":1398},[3182],{"type":55,"value":3183},"d",{"type":49,"tag":502,"props":3185,"children":3186},{"style":1404},[3187],{"type":55,"value":1781},{"type":49,"tag":502,"props":3189,"children":3190},{"style":1416},[3191],{"type":55,"value":3192}," --format-type",{"type":49,"tag":502,"props":3194,"children":3195},{"style":1416},[3196],{"type":55,"value":3197}," json",{"type":49,"tag":502,"props":3199,"children":3200},{"style":1398},[3201],{"type":55,"value":2329},{"type":49,"tag":502,"props":3203,"children":3204},{"class":504,"line":523},[3205,3210,3215,3220,3225],{"type":49,"tag":502,"props":3206,"children":3207},{"style":1404},[3208],{"type":55,"value":3209},"  |",{"type":49,"tag":502,"props":3211,"children":3212},{"style":1410},[3213],{"type":55,"value":3214}," jq",{"type":49,"tag":502,"props":3216,"children":3217},{"style":1404},[3218],{"type":55,"value":3219}," '",{"type":49,"tag":502,"props":3221,"children":3222},{"style":1416},[3223],{"type":55,"value":3224},"{status, tasks: [.groups[].tasks[] | {name, status, exit_code}]}",{"type":49,"tag":502,"props":3226,"children":3227},{"style":1404},[3228],{"type":55,"value":3229},"'\n",{"type":49,"tag":502,"props":3231,"children":3232},{"class":504,"line":2699},[3233],{"type":49,"tag":502,"props":3234,"children":3236},{"emptyLinePlaceholder":3235},true,[3237],{"type":55,"value":3238},"\n",{"type":49,"tag":502,"props":3240,"children":3241},{"class":504,"line":2731},[3242],{"type":49,"tag":502,"props":3243,"children":3244},{"style":3150},[3245],{"type":55,"value":3246},"# Logs for a specific task\n",{"type":49,"tag":502,"props":3248,"children":3249},{"class":504,"line":2743},[3250,3254,3258,3263,3267,3271,3275,3279,3284,3288,3293,3297,3301,3306],{"type":49,"tag":502,"props":3251,"children":3252},{"style":1410},[3253],{"type":55,"value":279},{"type":49,"tag":502,"props":3255,"children":3256},{"style":1416},[3257],{"type":55,"value":2314},{"type":49,"tag":502,"props":3259,"children":3260},{"style":1416},[3261],{"type":55,"value":3262}," logs",{"type":49,"tag":502,"props":3264,"children":3265},{"style":1404},[3266],{"type":55,"value":2710},{"type":49,"tag":502,"props":3268,"children":3269},{"style":1416},[3270],{"type":55,"value":3178},{"type":49,"tag":502,"props":3272,"children":3273},{"style":1398},[3274],{"type":55,"value":3183},{"type":49,"tag":502,"props":3276,"children":3277},{"style":1404},[3278],{"type":55,"value":1781},{"type":49,"tag":502,"props":3280,"children":3281},{"style":1416},[3282],{"type":55,"value":3283}," --task",{"type":49,"tag":502,"props":3285,"children":3286},{"style":1404},[3287],{"type":55,"value":2710},{"type":49,"tag":502,"props":3289,"children":3290},{"style":1416},[3291],{"type":55,"value":3292},"task_nam",{"type":49,"tag":502,"props":3294,"children":3295},{"style":1398},[3296],{"type":55,"value":1776},{"type":49,"tag":502,"props":3298,"children":3299},{"style":1404},[3300],{"type":55,"value":1781},{"type":49,"tag":502,"props":3302,"children":3303},{"style":1416},[3304],{"type":55,"value":3305}," -n",{"type":49,"tag":502,"props":3307,"children":3309},{"style":3308},"--shiki-light:#F76D47;--shiki-default:#F78C6C;--shiki-dark:#F78C6C",[3310],{"type":55,"value":3311}," 200\n",{"type":49,"tag":502,"props":3313,"children":3314},{"class":504,"line":2774},[3315],{"type":49,"tag":502,"props":3316,"children":3317},{"emptyLinePlaceholder":3235},[3318],{"type":55,"value":3238},{"type":49,"tag":502,"props":3320,"children":3321},{"class":504,"line":2788},[3322],{"type":49,"tag":502,"props":3323,"children":3324},{"style":3150},[3325],{"type":55,"value":3326},"# Output retrieval\n",{"type":49,"tag":502,"props":3328,"children":3329},{"class":504,"line":2817},[3330,3334,3339,3343,3348,3352,3357,3361],{"type":49,"tag":502,"props":3331,"children":3332},{"style":1410},[3333],{"type":55,"value":279},{"type":49,"tag":502,"props":3335,"children":3336},{"style":1416},[3337],{"type":55,"value":3338}," data",{"type":49,"tag":502,"props":3340,"children":3341},{"style":1416},[3342],{"type":55,"value":2201},{"type":49,"tag":502,"props":3344,"children":3345},{"style":1416},[3346],{"type":55,"value":3347}," --no-pager",{"type":49,"tag":502,"props":3349,"children":3350},{"style":1404},[3351],{"type":55,"value":2710},{"type":49,"tag":502,"props":3353,"children":3354},{"style":1416},[3355],{"type":55,"value":3356},"output_ur",{"type":49,"tag":502,"props":3358,"children":3359},{"style":1398},[3360],{"type":55,"value":2720},{"type":49,"tag":502,"props":3362,"children":3363},{"style":1404},[3364],{"type":55,"value":3365},">\n",{"type":49,"tag":502,"props":3367,"children":3368},{"class":504,"line":2848},[3369,3373,3377,3382,3386,3390,3394,3398,3402,3407,3412,3416],{"type":49,"tag":502,"props":3370,"children":3371},{"style":1410},[3372],{"type":55,"value":279},{"type":49,"tag":502,"props":3374,"children":3375},{"style":1416},[3376],{"type":55,"value":3338},{"type":49,"tag":502,"props":3378,"children":3379},{"style":1416},[3380],{"type":55,"value":3381}," download",{"type":49,"tag":502,"props":3383,"children":3384},{"style":1404},[3385],{"type":55,"value":2710},{"type":49,"tag":502,"props":3387,"children":3388},{"style":1416},[3389],{"type":55,"value":3356},{"type":49,"tag":502,"props":3391,"children":3392},{"style":1398},[3393],{"type":55,"value":2720},{"type":49,"tag":502,"props":3395,"children":3396},{"style":1404},[3397],{"type":55,"value":1781},{"type":49,"tag":502,"props":3399,"children":3400},{"style":1404},[3401],{"type":55,"value":2710},{"type":49,"tag":502,"props":3403,"children":3404},{"style":1416},[3405],{"type":55,"value":3406},"local_di",{"type":49,"tag":502,"props":3408,"children":3409},{"style":1398},[3410],{"type":55,"value":3411},"r",{"type":49,"tag":502,"props":3413,"children":3414},{"style":1404},[3415],{"type":55,"value":1781},{"type":49,"tag":502,"props":3417,"children":3418},{"style":1416},[3419],{"type":55,"value":3420},"\u002F\n",{"type":49,"tag":58,"props":3422,"children":3423},{},[3424],{"type":55,"value":3425},"For completion artifacts, always mirror the full run output into workspace:",{"type":49,"tag":491,"props":3427,"children":3429},{"className":1386,"code":3428,"language":1388,"meta":496,"style":496},"ROOT=\"$(git rev-parse --show-toplevel)\"\nRUN_LOCAL_DIR=\"$ROOT\u002Fmedia\u002Fvda\u002Fruns\u002F\u003Crun_id>\"\nmkdir -p \"$RUN_LOCAL_DIR\"\nosmo data download \"\u003Cstorage_url>\u002Fdatasets\u002F\u003Cdataset>-outputs\u002F\u003Crun_id>\u002F\" \"$RUN_LOCAL_DIR\u002F\"\n",[3430],{"type":49,"tag":92,"props":3431,"children":3432},{"__ignoreMap":496},[3433,3461,3491,3518],{"type":49,"tag":502,"props":3434,"children":3435},{"class":504,"line":505},[3436,3441,3445,3449,3453,3457],{"type":49,"tag":502,"props":3437,"children":3438},{"style":1398},[3439],{"type":55,"value":3440},"ROOT",{"type":49,"tag":502,"props":3442,"children":3443},{"style":1404},[3444],{"type":55,"value":1452},{"type":49,"tag":502,"props":3446,"children":3447},{"style":1404},[3448],{"type":55,"value":2566},{"type":49,"tag":502,"props":3450,"children":3451},{"style":1410},[3452],{"type":55,"value":2582},{"type":49,"tag":502,"props":3454,"children":3455},{"style":1416},[3456],{"type":55,"value":2587},{"type":49,"tag":502,"props":3458,"children":3459},{"style":1404},[3460],{"type":55,"value":2616},{"type":49,"tag":502,"props":3462,"children":3463},{"class":504,"line":514},[3464,3469,3473,3477,3482,3487],{"type":49,"tag":502,"props":3465,"children":3466},{"style":1398},[3467],{"type":55,"value":3468},"RUN_LOCAL_DIR",{"type":49,"tag":502,"props":3470,"children":3471},{"style":1404},[3472],{"type":55,"value":1452},{"type":49,"tag":502,"props":3474,"children":3475},{"style":1404},[3476],{"type":55,"value":1457},{"type":49,"tag":502,"props":3478,"children":3479},{"style":1398},[3480],{"type":55,"value":3481},"$ROOT",{"type":49,"tag":502,"props":3483,"children":3484},{"style":1416},[3485],{"type":55,"value":3486},"\u002Fmedia\u002Fvda\u002Fruns\u002F\u003Crun_id>",{"type":49,"tag":502,"props":3488,"children":3489},{"style":1404},[3490],{"type":55,"value":1472},{"type":49,"tag":502,"props":3492,"children":3493},{"class":504,"line":523},[3494,3499,3504,3509,3514],{"type":49,"tag":502,"props":3495,"children":3496},{"style":1410},[3497],{"type":55,"value":3498},"mkdir",{"type":49,"tag":502,"props":3500,"children":3501},{"style":1416},[3502],{"type":55,"value":3503}," -p",{"type":49,"tag":502,"props":3505,"children":3506},{"style":1404},[3507],{"type":55,"value":3508}," \"",{"type":49,"tag":502,"props":3510,"children":3511},{"style":1398},[3512],{"type":55,"value":3513},"$RUN_LOCAL_DIR",{"type":49,"tag":502,"props":3515,"children":3516},{"style":1404},[3517],{"type":55,"value":1472},{"type":49,"tag":502,"props":3519,"children":3520},{"class":504,"line":2699},[3521,3525,3529,3533,3537,3542,3546,3550,3554,3558],{"type":49,"tag":502,"props":3522,"children":3523},{"style":1410},[3524],{"type":55,"value":279},{"type":49,"tag":502,"props":3526,"children":3527},{"style":1416},[3528],{"type":55,"value":3338},{"type":49,"tag":502,"props":3530,"children":3531},{"style":1416},[3532],{"type":55,"value":3381},{"type":49,"tag":502,"props":3534,"children":3535},{"style":1404},[3536],{"type":55,"value":3508},{"type":49,"tag":502,"props":3538,"children":3539},{"style":1416},[3540],{"type":55,"value":3541},"\u003Cstorage_url>\u002Fdatasets\u002F\u003Cdataset>-outputs\u002F\u003Crun_id>\u002F",{"type":49,"tag":502,"props":3543,"children":3544},{"style":1404},[3545],{"type":55,"value":1457},{"type":49,"tag":502,"props":3547,"children":3548},{"style":1404},[3549],{"type":55,"value":3508},{"type":49,"tag":502,"props":3551,"children":3552},{"style":1398},[3553],{"type":55,"value":3513},{"type":49,"tag":502,"props":3555,"children":3556},{"style":1416},[3557],{"type":55,"value":183},{"type":49,"tag":502,"props":3559,"children":3560},{"style":1404},[3561],{"type":55,"value":1472},{"type":49,"tag":58,"props":3563,"children":3564},{},[3565,3567,3573],{"type":55,"value":3566},"For runs expected to exceed two minutes, send heartbeat updates at least every\ntwo minutes. For media evidence, emit one standalone ",{"type":49,"tag":92,"props":3568,"children":3570},{"className":3569},[],[3571],{"type":55,"value":3572},"MEDIA:\u003Cabsolute-path>",{"type":55,"value":3574},"\nline per message bubble.",{"type":49,"tag":58,"props":3576,"children":3577},{},[3578],{"type":55,"value":3579},"Execution continuity requirement:",{"type":49,"tag":1022,"props":3581,"children":3582},{},[3583,3588,3593,3598],{"type":49,"tag":410,"props":3584,"children":3585},{},[3586],{"type":55,"value":3587},"Heartbeats must report progress while continuing work; they are status updates,\nnot permission prompts.",{"type":49,"tag":410,"props":3589,"children":3590},{},[3591],{"type":55,"value":3592},"Do not stop between green stages waiting for approval.",{"type":49,"tag":410,"props":3594,"children":3595},{},[3596],{"type":55,"value":3597},"Pause only on blocking failures or explicit user stop\u002Fredirect.",{"type":49,"tag":410,"props":3599,"children":3600},{},[3601],{"type":55,"value":3602},"If submit fails on interpolation, rerun once with the same canonical single-flag\nshape and corrected values; do not loop through ad-hoc flag experiments.",{"type":49,"tag":58,"props":3604,"children":3605},{},[3606],{"type":55,"value":3607},"MEDIA formatting is strict:",{"type":49,"tag":1022,"props":3609,"children":3610},{},[3611,3622,3635,3640,3645],{"type":49,"tag":410,"props":3612,"children":3613},{},[3614,3616],{"type":55,"value":3615},"Emit exactly one line: ",{"type":49,"tag":92,"props":3617,"children":3619},{"className":3618},[],[3620],{"type":55,"value":3621},"MEDIA:\u002Fabsolute\u002Fpath\u002Fto\u002Ffile.mp4",{"type":49,"tag":410,"props":3623,"children":3624},{},[3625,3627,3633],{"type":55,"value":3626},"Keep ",{"type":49,"tag":92,"props":3628,"children":3630},{"className":3629},[],[3631],{"type":55,"value":3632},"MEDIA:",{"type":55,"value":3634}," contiguous on a single line (never split across lines).",{"type":49,"tag":410,"props":3636,"children":3637},{},[3638],{"type":55,"value":3639},"No extra text in the same bubble.",{"type":49,"tag":410,"props":3641,"children":3642},{},[3643],{"type":55,"value":3644},"No code fences, bullets, or quotes around the directive.",{"type":49,"tag":410,"props":3646,"children":3647},{},[3648],{"type":55,"value":3649},"If render fails: retry once from a stable workspace path, then emit PNG fallback.",{"type":49,"tag":64,"props":3651,"children":3653},{"id":3652},"post-run-comparison-evidence-required-for-augmented-flows",[3654],{"type":55,"value":3655},"Post-Run Comparison Evidence (required for augmented flows)",{"type":49,"tag":58,"props":3657,"children":3658},{},[3659,3661,3666,3667,3672,3674,3679],{"type":55,"value":3660},"Applies to ",{"type":49,"tag":92,"props":3662,"children":3664},{"className":3663},[],[3665],{"type":55,"value":427},{"type":55,"value":159},{"type":49,"tag":92,"props":3668,"children":3670},{"className":3669},[],[3671],{"type":55,"value":434},{"type":55,"value":3673},", and ",{"type":49,"tag":92,"props":3675,"children":3677},{"className":3676},[],[3678],{"type":55,"value":442},{"type":55,"value":3680}," after a\nsuccessful run.",{"type":49,"tag":58,"props":3682,"children":3683},{},[3684],{"type":55,"value":3685},"Required completion output (do not stop at raw output URLs):",{"type":49,"tag":406,"props":3687,"children":3688},{},[3689,3808,3897],{"type":49,"tag":410,"props":3690,"children":3691},{},[3692,3694],{"type":55,"value":3693},"Stage full outputs + input video into workspace-local path:",{"type":49,"tag":491,"props":3695,"children":3697},{"className":1386,"code":3696,"language":1388,"meta":496,"style":496},"bash scripts\u002Fstage_run_artifacts.sh \\\n  --storage-url \u003Cstorage_url> --dataset \u003Cdataset> --run-id \u003Crun_id> --video \u003Cvideo>\n",[3698],{"type":49,"tag":92,"props":3699,"children":3700},{"__ignoreMap":496},[3701,3717],{"type":49,"tag":502,"props":3702,"children":3703},{"class":504,"line":505},[3704,3708,3713],{"type":49,"tag":502,"props":3705,"children":3706},{"style":1410},[3707],{"type":55,"value":1388},{"type":49,"tag":502,"props":3709,"children":3710},{"style":1416},[3711],{"type":55,"value":3712}," scripts\u002Fstage_run_artifacts.sh",{"type":49,"tag":502,"props":3714,"children":3715},{"style":1398},[3716],{"type":55,"value":2329},{"type":49,"tag":502,"props":3718,"children":3719},{"class":504,"line":514},[3720,3725,3729,3734,3738,3742,3747,3751,3755,3759,3763,3768,3772,3777,3781,3785,3790,3794,3799,3804],{"type":49,"tag":502,"props":3721,"children":3722},{"style":1416},[3723],{"type":55,"value":3724},"  --storage-url",{"type":49,"tag":502,"props":3726,"children":3727},{"style":1404},[3728],{"type":55,"value":2710},{"type":49,"tag":502,"props":3730,"children":3731},{"style":1416},[3732],{"type":55,"value":3733},"storage_ur",{"type":49,"tag":502,"props":3735,"children":3736},{"style":1398},[3737],{"type":55,"value":2720},{"type":49,"tag":502,"props":3739,"children":3740},{"style":1404},[3741],{"type":55,"value":1781},{"type":49,"tag":502,"props":3743,"children":3744},{"style":1416},[3745],{"type":55,"value":3746}," --dataset",{"type":49,"tag":502,"props":3748,"children":3749},{"style":1404},[3750],{"type":55,"value":2710},{"type":49,"tag":502,"props":3752,"children":3753},{"style":1416},[3754],{"type":55,"value":2758},{"type":49,"tag":502,"props":3756,"children":3757},{"style":1398},[3758],{"type":55,"value":2763},{"type":49,"tag":502,"props":3760,"children":3761},{"style":1404},[3762],{"type":55,"value":1781},{"type":49,"tag":502,"props":3764,"children":3765},{"style":1416},[3766],{"type":55,"value":3767}," --run-id",{"type":49,"tag":502,"props":3769,"children":3770},{"style":1404},[3771],{"type":55,"value":2710},{"type":49,"tag":502,"props":3773,"children":3774},{"style":1416},[3775],{"type":55,"value":3776},"run_i",{"type":49,"tag":502,"props":3778,"children":3779},{"style":1398},[3780],{"type":55,"value":3183},{"type":49,"tag":502,"props":3782,"children":3783},{"style":1404},[3784],{"type":55,"value":1781},{"type":49,"tag":502,"props":3786,"children":3787},{"style":1416},[3788],{"type":55,"value":3789}," --video",{"type":49,"tag":502,"props":3791,"children":3792},{"style":1404},[3793],{"type":55,"value":2710},{"type":49,"tag":502,"props":3795,"children":3796},{"style":1416},[3797],{"type":55,"value":3798},"vide",{"type":49,"tag":502,"props":3800,"children":3801},{"style":1398},[3802],{"type":55,"value":3803},"o",{"type":49,"tag":502,"props":3805,"children":3806},{"style":1404},[3807],{"type":55,"value":3365},{"type":49,"tag":410,"props":3809,"children":3810},{},[3811,3813],{"type":55,"value":3812},"Render side-by-side from that local run copy:",{"type":49,"tag":491,"props":3814,"children":3816},{"className":1386,"code":3815,"language":1388,"meta":496,"style":496},"bash scripts\u002Frender_side_by_side.sh \\\n  --run-local-dir \"\u003Crepo>\u002Fmedia\u002Fvda\u002Fruns\u002F\u003Crun_id>\" --dataset \u003Cdataset> --video \u003Cvideo>\n",[3817],{"type":49,"tag":92,"props":3818,"children":3819},{"__ignoreMap":496},[3820,3836],{"type":49,"tag":502,"props":3821,"children":3822},{"class":504,"line":505},[3823,3827,3832],{"type":49,"tag":502,"props":3824,"children":3825},{"style":1410},[3826],{"type":55,"value":1388},{"type":49,"tag":502,"props":3828,"children":3829},{"style":1416},[3830],{"type":55,"value":3831}," scripts\u002Frender_side_by_side.sh",{"type":49,"tag":502,"props":3833,"children":3834},{"style":1398},[3835],{"type":55,"value":2329},{"type":49,"tag":502,"props":3837,"children":3838},{"class":504,"line":514},[3839,3844,3848,3853,3857,3861,3865,3869,3873,3877,3881,3885,3889,3893],{"type":49,"tag":502,"props":3840,"children":3841},{"style":1416},[3842],{"type":55,"value":3843},"  --run-local-dir",{"type":49,"tag":502,"props":3845,"children":3846},{"style":1404},[3847],{"type":55,"value":3508},{"type":49,"tag":502,"props":3849,"children":3850},{"style":1416},[3851],{"type":55,"value":3852},"\u003Crepo>\u002Fmedia\u002Fvda\u002Fruns\u002F\u003Crun_id>",{"type":49,"tag":502,"props":3854,"children":3855},{"style":1404},[3856],{"type":55,"value":1457},{"type":49,"tag":502,"props":3858,"children":3859},{"style":1416},[3860],{"type":55,"value":3746},{"type":49,"tag":502,"props":3862,"children":3863},{"style":1404},[3864],{"type":55,"value":2710},{"type":49,"tag":502,"props":3866,"children":3867},{"style":1416},[3868],{"type":55,"value":2758},{"type":49,"tag":502,"props":3870,"children":3871},{"style":1398},[3872],{"type":55,"value":2763},{"type":49,"tag":502,"props":3874,"children":3875},{"style":1404},[3876],{"type":55,"value":1781},{"type":49,"tag":502,"props":3878,"children":3879},{"style":1416},[3880],{"type":55,"value":3789},{"type":49,"tag":502,"props":3882,"children":3883},{"style":1404},[3884],{"type":55,"value":2710},{"type":49,"tag":502,"props":3886,"children":3887},{"style":1416},[3888],{"type":55,"value":3798},{"type":49,"tag":502,"props":3890,"children":3891},{"style":1398},[3892],{"type":55,"value":3803},{"type":49,"tag":502,"props":3894,"children":3895},{"style":1404},[3896],{"type":55,"value":3365},{"type":49,"tag":410,"props":3898,"children":3899},{},[3900,3902],{"type":55,"value":3901},"Emit MEDIA from the local run copy and include:",{"type":49,"tag":1022,"props":3903,"children":3904},{},[3905,3933,3944],{"type":49,"tag":410,"props":3906,"children":3907},{},[3908,3910,3916,3918,3924,3926,3932],{"type":55,"value":3909},"augmentation summary from ",{"type":49,"tag":92,"props":3911,"children":3913},{"className":3912},[],[3914],{"type":55,"value":3915},"\u003Crun_local_dir>\u002Fsetup_b0\u002Fconfigs\u002Fmanifest.yaml",{"type":55,"value":3917},"\n(",{"type":49,"tag":92,"props":3919,"children":3921},{"className":3920},[],[3922],{"type":55,"value":3923},"sampled_vars",{"type":55,"value":3925}," for ",{"type":49,"tag":92,"props":3927,"children":3929},{"className":3928},[],[3930],{"type":55,"value":3931},"\u003Cvideo>_aug0",{"type":55,"value":2592},{"type":49,"tag":410,"props":3934,"children":3935},{},[3936,3938],{"type":55,"value":3937},"auto-labeling summary from ",{"type":49,"tag":92,"props":3939,"children":3941},{"className":3940},[],[3942],{"type":55,"value":3943},"\u003Crun_local_dir>\u002Foutputs\u002Fpseudo_labeled_augmented\u002F\u003Cvideo>_aug0",{"type":49,"tag":410,"props":3945,"children":3946},{},[3947,3949,3954,3955,3960,3962],{"type":55,"value":3948},"for ",{"type":49,"tag":92,"props":3950,"children":3952},{"className":3951},[],[3953],{"type":55,"value":434},{"type":55,"value":371},{"type":49,"tag":92,"props":3956,"children":3958},{"className":3957},[],[3959],{"type":55,"value":442},{"type":55,"value":3961},", original-label summary from\n",{"type":49,"tag":92,"props":3963,"children":3965},{"className":3964},[],[3966],{"type":55,"value":3967},"\u003Crun_local_dir>\u002Foutputs\u002Fpseudo_labeled\u002F\u003Cvideo>",{"type":49,"tag":58,"props":3969,"children":3970},{},[3971,3972,3978],{"type":55,"value":2287},{"type":49,"tag":92,"props":3973,"children":3975},{"className":3974},[],[3976],{"type":55,"value":3977},"ffmpeg",{"type":55,"value":3979}," is unavailable, emit input and augmented MEDIA from the same local\nrun copy and still provide augmentation + auto-labeling summaries.",{"type":49,"tag":58,"props":3981,"children":3982},{},[3983,3985,3990],{"type":55,"value":3984},"For demo runs (no user video provided), explicitly state that input came from\n",{"type":49,"tag":92,"props":3986,"children":3988},{"className":3987},[],[3989],{"type":55,"value":1262},{"type":55,"value":107},{"type":49,"tag":64,"props":3992,"children":3994},{"id":3993},"supporting-files",[3995],{"type":55,"value":3996},"Supporting files",{"type":49,"tag":58,"props":3998,"children":3999},{},[4000],{"type":55,"value":4001},"Use these canonical locations:",{"type":49,"tag":1022,"props":4003,"children":4004},{},[4005,4016,4034,4045,4063,4080],{"type":49,"tag":410,"props":4006,"children":4007},{},[4008,4010],{"type":55,"value":4009},"Workflows: ",{"type":49,"tag":92,"props":4011,"children":4013},{"className":4012},[],[4014],{"type":55,"value":4015},"assets\u002Fconfigs\u002Fosmo\u002F*.yaml",{"type":49,"tag":410,"props":4017,"children":4018},{},[4019,4021,4027,4028],{"type":55,"value":4020},"Runtime scripts: ",{"type":49,"tag":92,"props":4022,"children":4024},{"className":4023},[],[4025],{"type":55,"value":4026},"scripts\u002F*.sh",{"type":55,"value":159},{"type":49,"tag":92,"props":4029,"children":4031},{"className":4030},[],[4032],{"type":55,"value":4033},"scripts\u002F*.py",{"type":49,"tag":410,"props":4035,"children":4036},{},[4037,4039],{"type":55,"value":4038},"Flow walkthroughs: ",{"type":49,"tag":92,"props":4040,"children":4042},{"className":4041},[],[4043],{"type":55,"value":4044},"references\u002Fflows\u002F*.md",{"type":49,"tag":410,"props":4046,"children":4047},{},[4048,4050,4056,4057],{"type":55,"value":4049},"Setup and triage: ",{"type":49,"tag":92,"props":4051,"children":4053},{"className":4052},[],[4054],{"type":55,"value":4055},"references\u002Fsetup.md",{"type":55,"value":159},{"type":49,"tag":92,"props":4058,"children":4060},{"className":4059},[],[4061],{"type":55,"value":4062},"references\u002Ftroubleshooting.md",{"type":49,"tag":410,"props":4064,"children":4065},{},[4066,4068,4074,4075],{"type":55,"value":4067},"Images and endpoint policy: ",{"type":49,"tag":92,"props":4069,"children":4071},{"className":4070},[],[4072],{"type":55,"value":4073},"references\u002Fcontainer-images.md",{"type":55,"value":159},{"type":49,"tag":92,"props":4076,"children":4078},{"className":4077},[],[4079],{"type":55,"value":2152},{"type":49,"tag":410,"props":4081,"children":4082},{},[4083,4085],{"type":55,"value":4084},"Cookbook tuning: ",{"type":49,"tag":92,"props":4086,"children":4088},{"className":4087},[],[4089],{"type":55,"value":4090},"assets\u002Fcookbooks\u002FTUNING_GUIDE.md",{"type":49,"tag":4092,"props":4093,"children":4094},"style",{},[4095],{"type":55,"value":4096},"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":4098,"total":4252},[4099,4117,4134,4145,4157,4171,4184,4196,4209,4220,4234,4243],{"slug":4100,"name":4100,"fn":4101,"description":4102,"org":4103,"tags":4104,"stars":4114,"repoUrl":4115,"updatedAt":4116},"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},[4105,4108,4111],{"name":4106,"slug":4107,"type":15},"Documentation","documentation",{"name":4109,"slug":4110,"type":15},"MCP","mcp",{"name":4112,"slug":4113,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":4118,"name":4118,"fn":4119,"description":4120,"org":4121,"tags":4122,"stars":4131,"repoUrl":4132,"updatedAt":4133},"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},[4123,4126,4129],{"name":4124,"slug":4125,"type":15},"Containers","containers",{"name":4127,"slug":4128,"type":15},"Deployment","deployment",{"name":4130,"slug":495,"type":15},"Python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":4135,"name":4135,"fn":4136,"description":4137,"org":4138,"tags":4139,"stars":4131,"repoUrl":4132,"updatedAt":4144},"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},[4140,4143],{"name":4141,"slug":4142,"type":15},"CI\u002FCD","ci-cd",{"name":4127,"slug":4128,"type":15},"2026-07-14T05:25:59.97109",{"slug":4146,"name":4146,"fn":4147,"description":4148,"org":4149,"tags":4150,"stars":4131,"repoUrl":4132,"updatedAt":4156},"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},[4151,4152,4153],{"name":4141,"slug":4142,"type":15},{"name":4127,"slug":4128,"type":15},{"name":4154,"slug":4155,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":4158,"name":4158,"fn":4159,"description":4160,"org":4161,"tags":4162,"stars":4131,"repoUrl":4132,"updatedAt":4170},"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},[4163,4166,4167],{"name":4164,"slug":4165,"type":15},"Debugging","debugging",{"name":4154,"slug":4155,"type":15},{"name":4168,"slug":4169,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":4172,"name":4172,"fn":4173,"description":4174,"org":4175,"tags":4176,"stars":4131,"repoUrl":4132,"updatedAt":4183},"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},[4177,4180],{"name":4178,"slug":4179,"type":15},"Best Practices","best-practices",{"name":4181,"slug":4182,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":4185,"name":4185,"fn":4186,"description":4187,"org":4188,"tags":4189,"stars":4131,"repoUrl":4132,"updatedAt":4195},"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},[4190,4191,4194],{"name":17,"slug":18,"type":15},{"name":4192,"slug":4193,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":4197,"name":4197,"fn":4198,"description":4199,"org":4200,"tags":4201,"stars":4131,"repoUrl":4132,"updatedAt":4208},"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},[4202,4205],{"name":4203,"slug":4204,"type":15},"QA","qa",{"name":4206,"slug":4207,"type":15},"Testing","testing","2026-07-14T05:25:53.673039",{"slug":4210,"name":4210,"fn":4211,"description":4212,"org":4213,"tags":4214,"stars":4131,"repoUrl":4132,"updatedAt":4219},"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},[4215,4216],{"name":4127,"slug":4128,"type":15},{"name":4217,"slug":4218,"type":15},"Infrastructure","infrastructure","2026-07-14T05:25:49.362534",{"slug":4221,"name":4221,"fn":4222,"description":4223,"org":4224,"tags":4225,"stars":4131,"repoUrl":4132,"updatedAt":4233},"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},[4226,4229,4230],{"name":4227,"slug":4228,"type":15},"Code Review","code-review",{"name":4154,"slug":4155,"type":15},{"name":4231,"slug":4232,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":4235,"name":4235,"fn":4236,"description":4237,"org":4238,"tags":4239,"stars":4131,"repoUrl":4132,"updatedAt":4242},"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},[4240,4241],{"name":4203,"slug":4204,"type":15},{"name":4206,"slug":4207,"type":15},"2026-07-14T05:25:54.928983",{"slug":4244,"name":4244,"fn":4245,"description":4246,"org":4247,"tags":4248,"stars":4131,"repoUrl":4132,"updatedAt":4251},"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},[4249,4250],{"name":13,"slug":14,"type":15},{"name":4141,"slug":4142,"type":15},"2026-07-30T05:29:03.275638",496,{"items":4254,"total":4348},[4255,4272,4282,4296,4306,4319,4334],{"slug":4256,"name":4256,"fn":4257,"description":4258,"org":4259,"tags":4260,"stars":23,"repoUrl":24,"updatedAt":4271},"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},[4261,4264,4267,4268],{"name":4262,"slug":4263,"type":15},"Data Analysis","data-analysis",{"name":4265,"slug":4266,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":4269,"slug":4270,"type":15},"Performance","performance","2026-07-14T05:28:43.176466",{"slug":4273,"name":4273,"fn":4274,"description":4275,"org":4276,"tags":4277,"stars":23,"repoUrl":24,"updatedAt":4281},"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},[4278,4279,4280],{"name":4127,"slug":4128,"type":15},{"name":4217,"slug":4218,"type":15},{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":4283,"name":4283,"fn":4284,"description":4285,"org":4286,"tags":4287,"stars":23,"repoUrl":24,"updatedAt":4295},"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},[4288,4291,4292],{"name":4289,"slug":4290,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":4293,"slug":4294,"type":15},"Research","research","2026-07-14T05:28:06.816956",{"slug":4297,"name":4297,"fn":4298,"description":4299,"org":4300,"tags":4301,"stars":23,"repoUrl":24,"updatedAt":4305},"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},[4302,4303,4304],{"name":4262,"slug":4263,"type":15},{"name":9,"slug":8,"type":15},{"name":4206,"slug":4207,"type":15},"2026-07-17T05:29:03.913266",{"slug":4307,"name":4307,"fn":4308,"description":4309,"org":4310,"tags":4311,"stars":23,"repoUrl":24,"updatedAt":4318},"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},[4312,4313,4316,4317],{"name":13,"slug":14,"type":15},{"name":4314,"slug":4315,"type":15},"Imaging","imaging",{"name":9,"slug":8,"type":15},{"name":21,"slug":22,"type":15},"2026-07-17T05:28:53.905004",{"slug":4320,"name":4320,"fn":4321,"description":4322,"org":4323,"tags":4324,"stars":23,"repoUrl":24,"updatedAt":4333},"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},[4325,4326,4329,4330],{"name":4127,"slug":4128,"type":15},{"name":4327,"slug":4328,"type":15},"Docker","docker",{"name":9,"slug":8,"type":15},{"name":4331,"slug":4332,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":4335,"name":4335,"fn":4336,"description":4337,"org":4338,"tags":4339,"stars":23,"repoUrl":24,"updatedAt":4347},"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},[4340,4341,4344],{"name":9,"slug":8,"type":15},{"name":4342,"slug":4343,"type":15},"Quantum Computing","quantum-computing",{"name":4345,"slug":4346,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305]