[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6PKHQ90EFFGDC0SHR7GRRC7":3,"skill-same:01M6PKHQ90EFFGDC0SHR7GRRC7":178},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":39,"linkOnly":40,"markdownContent":41,"markdownBody":42,"frontMatter":43,"refName":50,"commitSha":51,"skillDirectory":52,"forksCount":14,"repositoryHtmlUrl":53,"repositoryHomepageUrl":6,"ownerHtmlUrl":54,"ownerBlogUrl":55,"ownerDescription":6,"updatedAt":56,"install":57,"resources":70},"01M6PKHQ90EFFGDC0SHR7GRRC7","tao-train-depth-anything-v2",null,"Monocular depth estimation using Metric Depth Anything v2 or Relative Depth Anything architectures. Predicts per-pixel depth from single RGB images. Use when training, evaluating, exporting, or running inference for a TAO monocular depth model. Trigger phrases include \"train monocular depth\", \"DepthAnything v2\", \"metric depth from single image\", \"monocular depth estimation\".","headmktlucca-ui\u002Fneuroads-lp","headmktlucca-ui","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F247379838?v=4","01M6PDVAV6M1W96D1SR1PDF8CS","NeuroAds","01M6PDVAZ4P281GNDKD45CYTF5",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38],"train","tao train depth anything v2","depth","tao","Bash","model","read","container","images","pixel","image","estimation","trigger","architectures","for","metric","NVIDIA","training","monocular","docker","2026-11-10 10:12:30",false,"---\nname: tao-train-depth-anything-v2\ndescription: Monocular depth estimation using Metric Depth Anything v2 or Relative Depth Anything architectures. Predicts\n  per-pixel depth from single RGB images. Use when training, evaluating, exporting, or running inference for a TAO\n  monocular depth model. Trigger phrases include \"train monocular depth\", \"DepthAnything v2\", \"metric depth from single\n  image\", \"monocular depth estimation\".\nlicense: Apache-2.0\ncompatibility: Requires docker + nvidia-container-toolkit.\nmetadata:\n  version: \"0.1.0\"\n  author: NVIDIA Corporation\nallowed-tools: Read Bash\ntags:\n- monocular\n- depth\n- estimation\n---\n\n# Depth Net Mono\n\nMonocular depth estimation using Metric Depth Anything v2 or Relative Depth Anything architectures. Predicts per-pixel depth from single RGB images.\n\nPretrained checkpoint loading varies by model variant and use case — see the **Pretrained checkpoint loading — use case matrix** in `references\u002Fparameters.md`.\n\nThe mono and stereo skills both invoke the unified TAO `depth_net` CLI inside the container; the mono\u002Fstereo family is selected via `model.model_type` (full parameter glossary in `references\u002Fparameters.md`).\n\nFor TAO Deploy TensorRT actions (`gen_trt_engine`, TensorRT `evaluate`, and TensorRT `inference`), read `references\u002Ftao-deploy-depth-anything-v2.md` first. The deploy spec template lives in this skill's `references\u002Fspec_template_deploy.yaml`.\n\n## Train Action Policy\n\nThis model is AutoML-enabled at the model layer. Before handling any train-stage request, read `references\u002Fskill_info.yaml` and resolve the run override from either an explicit `automl_policy` value or the user's workflow request. Treat phrases like \"turn off AutoML\", \"disable AutoML\", \"no HPO\", or \"plain training\" as `automl_policy: off` for this run only; otherwise default to `auto`. When `automl_policy: auto`, `automl_enabled: true`, and both `schemas\u002Ftrain.schema.json` and `references\u002Fspec_template_train.yaml` are packaged, route the train action through `tao-skill-bank:tao-run-automl` by default with this model's `skill_dir`. Preserve workflow\u002Fapplication overrides for datasets, specs, output directories, GPU\u002Fplatform settings, parent checkpoints, and `automl_policy`. Use direct model training only when `automl_policy: off` or the packaged train schema\u002Ftemplate is missing; in the missing-schema case, report that AutoML is enabled but not runnable for this model until schemas are generated.\n\nNon-train actions such as `evaluate`, `inference`, `export`, and deploy flows stay in this model skill. The per-run `automl_policy` override does not change model metadata.\n\n## Workflow\n\n### Prerequisites — data accessibility\n\nYour dataset (RGB images + GT depth files) must be reachable from inside the container:\n- **SDK runner**: place files at the S3 paths the runner resolves (the `S3_TRAIN` \u002F `S3_EVAL` placeholders shown in the spec overrides). The runner handles S3 → container-path mounting transparently.\n- **Direct `docker run`** (e.g. local testing): mount the host dataset root read-only at the same in-container path:\n\n```\ndocker run ... -v \u003Chost_data_root>:\u003Chost_data_root>:ro \u003Ccontainer> ...\n```\n\nThe same accessibility requirement applies to the `\u003Coutput_dir>` written by all actions.\n\n### Step 1 — Annotation file\n\nPer-line annotation file referenced by `data_sources[*].data_file`:\n\n| Columns | Format | Use |\n|---|---|---|\n| 1 | `\u003Cimage>` | Mono inference (no GT) |\n| 2 | `\u003Cimage> \u003Cgt_depth>` | Mono with GT |\n\nIf you already have one, point to it. Otherwise generate via `depth_net convert`:\n\n```\ndepth_net convert -e \u003Cconvert_spec.yaml>\n```\n\n`convert_spec.yaml` template:\n\n```yaml\ndata_root: \u003Cdirectory whose immediate children are scene\u002Fsample folders that contain your image+depth files; convert walks data_root recursively but expects per-scene subdirectories at one level below>\nimage_dir_pattern: [\u003Csubstring matching left\u002FRGB image paths>]\ndepth_dir_pattern: [\u003Csubstring matching GT depth paths>]\nimage_extension: ''     # optional .endswith filter, e.g. '.jpg'\ndepth_extension: ''     # optional, swapped during depth derivation, e.g. '.png'\nsplit_ratio: 0.0        # 0.0\u002F1.0 = test-only; 0.8 = 80\u002F20 train+val\n```\n\n`convert` walks `data_root` recursively, selects paths whose path-string contains *all* substrings in `image_dir_pattern` (AND-filter), then derives the depth path by replacing `image_dir_pattern[0]` with `depth_dir_pattern[0]` and `image_extension` with `depth_extension`. Inspect your dataset's directory layout and identify the substring distinguishing RGB images from depth files (e.g. `rgb_` vs `sync_depth_`).\n\n`data_root` must point at the parent that contains the per-scene subdirectories (e.g. for NYU eval, use `\u002Fdata\u002Fnyu_v2\u002Feval\u002Ftest`, not `\u002Fdata\u002Fnyu_v2\u002Feval\u002Ftest\u002Fbathroom` — the latter limits the walk to a single scene). Always include the leading dot in `image_extension` \u002F `depth_extension` (e.g. `'.jpg'` not `'jpg'`); the substring swap is form-sensitive and a mismatch silently corrupts derived paths.\n\n### Step 2 — Pair `model_type` and `dataset_name` based on your data\n\nDefault — generic class for each task:\n\n| Data category | `model_type` | `dataset_name` |\n|---|---|---|\n| Disparity-encoded data (pixels) | `RelativeDepthAnything` | `RelativeMonoDataset` |\n| Metric depth (meters) | `MetricDepthAnything` | `MetricMonoDataset` |\n| Mono inference (no GT, any image) | matches train choice | `RelativeMonoDataset` or `MetricMonoDataset` |\n\nDataset-specific class — switch when the data needs preprocessing the generic class does not perform:\n\n| Special case | `model_type` | `dataset_name` | What the class adds |\n|---|---|---|---|\n| NYU `sync_depth_*.png` (raw uint16 millimetres) — relative | `RelativeDepthAnything` | `NYUDV2Relative` | mm→m unit conversion + Eigen evaluation crop |\n| NYU `sync_depth_*.png` (raw uint16 millimetres) — metric | `MetricDepthAnything` | `NYUDV2` | same |\n\nUsing a generic class on data that requires unit conversion (e.g. raw NYU uint16 PNGs) results in an empty valid mask and silent `train_loss = NaN`. Match the class to your data's encoding.\n\n### Step 3 — Write spec yaml from the spec overrides\n\nCopy the action block from `references\u002Fspec-overrides.md`. Replace:\n- `model.model_type` from Step 2\n- `dataset.\u003C...>.data_sources[*].dataset_name` from Step 2\n- `data_sources[*].data_file` with the path from Step 1 (S3 path under SDK runner, host path for direct docker)\n- For metric finetune: additionally apply the **Metric Variant Finetuning Recipe** in `references\u002Ffinetuning.md`.\n\nFor mono training set `train.precision: fp32` (recommended) or `bf16` (Ampere SM80+, alternative).\n\n### Step 4 — Run\n\n```\ndocker run --gpus 'device=0' --shm-size 16G --ipc=host \\\n  --user $(id -u):$(id -g) \\\n  -v \u003Cdata_root>:\u003Cdata_root>:ro \\\n  -v \u003Coutput_dir>:\u003Coutput_dir> \\\n  \u003Ccontainer> \\\n  depth_net \u003Caction> -e \u003Cspec.yaml>\n```\n\nWithout `--user $(id -u):$(id -g)` the container writes outputs as `nobody:nogroup`, blocking host-side cleanup and retry.\n\n### Step 5 — Verify\n\n- Container exit code 0\n- `status.json` `kpi` block populated\n- For `train`: inspect per-step `train_loss` directly — the entrypoint reports `Execution status: PASS` even when `train_loss = NaN` (see the **Sanity-run PASS criteria** in `references\u002Ffinetuning.md`)\n- For `evaluate` \u002F `inference`: artifacts under `results_dir`\n\nFor TAO Deploy TensorRT actions (`gen_trt_engine`, TensorRT `evaluate`, and TensorRT `inference`), read `references\u002Ftao-deploy-depth-anything-v2.md` first. Deploy spec templates live in this skill's `references\u002F` folder with the `spec_template_deploy_*.yaml` prefix.\n\n## Training Requirements\n\n- **Valid `dataset_name` values for mono `data_sources`** (case-insensitive): `ThreeDVLM`, `FSD`, `NvCLIP`, `IssacStereo`, `Crestereo`, `Middlebury`, `NYUDV2`, `NYUDV2Relative`, `RelativeMonoDataset`, `MetricMonoDataset`. `NYUDV2` carries metric depth GT (meters) — pair with `MetricDepthAnything`; `NYUDV2Relative` is the same data with relative-depth conventions — pair with `RelativeDepthAnything`.\n- **Monitoring metric:** val\u002Floss\n\n### Per-Action Dataset Requirements\n\n| Action | Spec Key | Source | Files | List? |\n|---|---|---|---|---|\n| evaluate | dataset.test_dataset.data_sources | eval_dataset | data_file: annotations.txt + dataset_name | Yes |\n| inference | dataset.infer_dataset.data_sources | inference_dataset | data_file: annotations.txt + dataset_name | Yes |\n| quantize | dataset.train_dataset.data_sources | train_datasets | data_file: annotations.txt + dataset_name | Yes |\n| quantize | dataset.val_dataset.data_sources | eval_dataset | data_file: annotations.txt + dataset_name | Yes |\n| quantize | dataset.quant_calibration_dataset.images_dir | train_datasets | images.tar.gz | No |\n| train | dataset.train_dataset.data_sources | train_datasets | data_file: annotations.txt + dataset_name | Yes |\n| train | dataset.val_dataset.data_sources | eval_dataset | data_file: annotations.txt + dataset_name | Yes |\n\n### Spec Overrides\n\nData source overrides are **mandatory for every action** — construct the data source paths from the Per-Action Dataset Requirements table above and include them in `spec_overrides`; each `data_sources` entry is a dict with the two mandatory fields `data_file` and `dataset_name`. See `references\u002Fspec-overrides.md` for the full per-action `train` \u002F `evaluate` \u002F `export` \u002F `inference` \u002F `quantize` override blocks and the precision recommendations.\n\n## Eval Dataset\n\nOptional. Val dataset configured via `dataset.val_dataset.data_sources` (each entry needs `data_file` and `dataset_name`).\n\n## Important Parameters\n\nFull parameter glossary (`model.*`, `train.*`, `dataset.*`, `export.*`, `inference.*` fields with options, defaults, and sources) plus the **Pretrained checkpoint loading — use case matrix** live in `references\u002Fparameters.md`. Key starting points: `model.model_type` (default `MetricDepthAnything`), `model.encoder` (default `vitl`), `train.optim.lr` (default 1e-4, AdamW), `train.precision` (`fp32` recommended), `dataset.{train,val,test,infer}_dataset.augmentation.crop_size` (default `[518, 518]`).\n\n## Finetuning Recipes\n\nRelative and Metric variant finetuning recipes — including required spec keys, the metric `dataset.{normalize_depth, min_depth, max_depth}` block required in both train AND export specs, trainer-enforced defaults (`clip_grad_norm: 0.1`, `warmup_steps: 20`, `weight_decay: 1e-4`), sanity-run overrides, and the **Sanity-run PASS criteria** for catching silent `train_loss = NaN` — are in `references\u002Ffinetuning.md`. Both recipes use `train.optim.lr: 5e-6` with `LambdaLR` (the AdamW default `1e-4` is too aggressive when finetuning from a converged\u002Fpretrained backbone).\n\n## Multi-GPU \u002F Multi-Node\n\n**Launch method:** Lightning-managed (single `python` process, Lightning spawns workers).\n\n| Spec Key | Description | Default |\n|----------|-------------|---------|\n| `train.num_gpus` | Number of GPUs | 1 |\n| `train.gpu_ids` | GPU device indices | [0] |\n| `train.num_nodes` | Number of nodes | 1 |\n| `train.distributed_strategy` | `ddp` or `fsdp` | `ddp` |\n\n- `ddp` with activation checkpointing: `find_unused_parameters=False`\n- `ddp` without: `find_unused_parameters=True`\n- `fsdp` forces precision to FP16\n\n**Multi-node env vars** (set by orchestrator): `WORLD_SIZE`, `NODE_RANK`, `MASTER_ADDR`, `MASTER_PORT`, `NUM_GPU_PER_NODE`.\n\n## Export \u002F TRT Defaults\n\n- TRT data types: FP32, BF16 (Ampere SM80+). FP16 is not supported for the ViT-L mono backbone.\n- Recommended TRT precision: `bf16`. Use `fp32` if BF16 hardware is unavailable.\n\nFull TAO Deploy reference: [tao-deploy-depth-anything-v2](references\u002Ftao-deploy-depth-anything-v2.md).\n\n## Hardware\n\nMinimum 1 GPU(s), recommended 2 GPU(s). 24GB+ VRAM per GPU. ViT-Large encoder is memory intensive. Use `fp32` (recommended) or `bf16` (Ampere SM80+, alternative) for training. Activation checkpointing is available for larger inputs.\n\n## Error Patterns\n\nCommon failure signatures and fixes — depth range mismatch, missing pretrained weights, `Key 'encoder' not in 'MonoBackBone'`, missing `dataset_name`, `depth_net_mono: not found`, metric variant hyperparameter sourcing, and the export refuse-to-overwrite ONNX error — are documented in `references\u002Ftroubleshooting.md`.\n\n## Spec Param \u002F Parent Model Inference\n\nModel-specific inference mappings (the full `depth_net_mono.config.json` per-action spec-field → inference-function table, plus `parent_model` \u002F `parent_job_id` resolution guidance) are in `references\u002Fspec-param-inference.md`. These mappings belong in MD, not in `config.json`; generated runners should read that reference and apply the mappings with SDK helpers before `create_job()`.\n","# Depth Net Mono\n\nMonocular depth estimation using Metric Depth Anything v2 or Relative Depth Anything architectures. Predicts per-pixel depth from single RGB images.\n\nPretrained checkpoint loading varies by model variant and use case — see the **Pretrained checkpoint loading — use case matrix** in `references\u002Fparameters.md`.\n\nThe mono and stereo skills both invoke the unified TAO `depth_net` CLI inside the container; the mono\u002Fstereo family is selected via `model.model_type` (full parameter glossary in `references\u002Fparameters.md`).\n\nFor TAO Deploy TensorRT actions (`gen_trt_engine`, TensorRT `evaluate`, and TensorRT `inference`), read `references\u002Ftao-deploy-depth-anything-v2.md` first. The deploy spec template lives in this skill's `references\u002Fspec_template_deploy.yaml`.\n\n## Train Action Policy\n\nThis model is AutoML-enabled at the model layer. Before handling any train-stage request, read `references\u002Fskill_info.yaml` and resolve the run override from either an explicit `automl_policy` value or the user's workflow request. Treat phrases like \"turn off AutoML\", \"disable AutoML\", \"no HPO\", or \"plain training\" as `automl_policy: off` for this run only; otherwise default to `auto`. When `automl_policy: auto`, `automl_enabled: true`, and both `schemas\u002Ftrain.schema.json` and `references\u002Fspec_template_train.yaml` are packaged, route the train action through `tao-skill-bank:tao-run-automl` by default with this model's `skill_dir`. Preserve workflow\u002Fapplication overrides for datasets, specs, output directories, GPU\u002Fplatform settings, parent checkpoints, and `automl_policy`. Use direct model training only when `automl_policy: off` or the packaged train schema\u002Ftemplate is missing; in the missing-schema case, report that AutoML is enabled but not runnable for this model until schemas are generated.\n\nNon-train actions such as `evaluate`, `inference`, `export`, and deploy flows stay in this model skill. The per-run `automl_policy` override does not change model metadata.\n\n## Workflow\n\n### Prerequisites — data accessibility\n\nYour dataset (RGB images + GT depth files) must be reachable from inside the container:\n- **SDK runner**: place files at the S3 paths the runner resolves (the `S3_TRAIN` \u002F `S3_EVAL` placeholders shown in the spec overrides). The runner handles S3 → container-path mounting transparently.\n- **Direct `docker run`** (e.g. local testing): mount the host dataset root read-only at the same in-container path:\n\n```\ndocker run ... -v \u003Chost_data_root>:\u003Chost_data_root>:ro \u003Ccontainer> ...\n```\n\nThe same accessibility requirement applies to the `\u003Coutput_dir>` written by all actions.\n\n### Step 1 — Annotation file\n\nPer-line annotation file referenced by `data_sources[*].data_file`:\n\n| Columns | Format | Use |\n|---|---|---|\n| 1 | `\u003Cimage>` | Mono inference (no GT) |\n| 2 | `\u003Cimage> \u003Cgt_depth>` | Mono with GT |\n\nIf you already have one, point to it. Otherwise generate via `depth_net convert`:\n\n```\ndepth_net convert -e \u003Cconvert_spec.yaml>\n```\n\n`convert_spec.yaml` template:\n\n```yaml\ndata_root: \u003Cdirectory whose immediate children are scene\u002Fsample folders that contain your image+depth files; convert walks data_root recursively but expects per-scene subdirectories at one level below>\nimage_dir_pattern: [\u003Csubstring matching left\u002FRGB image paths>]\ndepth_dir_pattern: [\u003Csubstring matching GT depth paths>]\nimage_extension: ''     # optional .endswith filter, e.g. '.jpg'\ndepth_extension: ''     # optional, swapped during depth derivation, e.g. '.png'\nsplit_ratio: 0.0        # 0.0\u002F1.0 = test-only; 0.8 = 80\u002F20 train+val\n```\n\n`convert` walks `data_root` recursively, selects paths whose path-string contains *all* substrings in `image_dir_pattern` (AND-filter), then derives the depth path by replacing `image_dir_pattern[0]` with `depth_dir_pattern[0]` and `image_extension` with `depth_extension`. Inspect your dataset's directory layout and identify the substring distinguishing RGB images from depth files (e.g. `rgb_` vs `sync_depth_`).\n\n`data_root` must point at the parent that contains the per-scene subdirectories (e.g. for NYU eval, use `\u002Fdata\u002Fnyu_v2\u002Feval\u002Ftest`, not `\u002Fdata\u002Fnyu_v2\u002Feval\u002Ftest\u002Fbathroom` — the latter limits the walk to a single scene). Always include the leading dot in `image_extension` \u002F `depth_extension` (e.g. `'.jpg'` not `'jpg'`); the substring swap is form-sensitive and a mismatch silently corrupts derived paths.\n\n### Step 2 — Pair `model_type` and `dataset_name` based on your data\n\nDefault — generic class for each task:\n\n| Data category | `model_type` | `dataset_name` |\n|---|---|---|\n| Disparity-encoded data (pixels) | `RelativeDepthAnything` | `RelativeMonoDataset` |\n| Metric depth (meters) | `MetricDepthAnything` | `MetricMonoDataset` |\n| Mono inference (no GT, any image) | matches train choice | `RelativeMonoDataset` or `MetricMonoDataset` |\n\nDataset-specific class — switch when the data needs preprocessing the generic class does not perform:\n\n| Special case | `model_type` | `dataset_name` | What the class adds |\n|---|---|---|---|\n| NYU `sync_depth_*.png` (raw uint16 millimetres) — relative | `RelativeDepthAnything` | `NYUDV2Relative` | mm→m unit conversion + Eigen evaluation crop |\n| NYU `sync_depth_*.png` (raw uint16 millimetres) — metric | `MetricDepthAnything` | `NYUDV2` | same |\n\nUsing a generic class on data that requires unit conversion (e.g. raw NYU uint16 PNGs) results in an empty valid mask and silent `train_loss = NaN`. Match the class to your data's encoding.\n\n### Step 3 — Write spec yaml from the spec overrides\n\nCopy the action block from `references\u002Fspec-overrides.md`. Replace:\n- `model.model_type` from Step 2\n- `dataset.\u003C...>.data_sources[*].dataset_name` from Step 2\n- `data_sources[*].data_file` with the path from Step 1 (S3 path under SDK runner, host path for direct docker)\n- For metric finetune: additionally apply the **Metric Variant Finetuning Recipe** in `references\u002Ffinetuning.md`.\n\nFor mono training set `train.precision: fp32` (recommended) or `bf16` (Ampere SM80+, alternative).\n\n### Step 4 — Run\n\n```\ndocker run --gpus 'device=0' --shm-size 16G --ipc=host \\\n  --user $(id -u):$(id -g) \\\n  -v \u003Cdata_root>:\u003Cdata_root>:ro \\\n  -v \u003Coutput_dir>:\u003Coutput_dir> \\\n  \u003Ccontainer> \\\n  depth_net \u003Caction> -e \u003Cspec.yaml>\n```\n\nWithout `--user $(id -u):$(id -g)` the container writes outputs as `nobody:nogroup`, blocking host-side cleanup and retry.\n\n### Step 5 — Verify\n\n- Container exit code 0\n- `status.json` `kpi` block populated\n- For `train`: inspect per-step `train_loss` directly — the entrypoint reports `Execution status: PASS` even when `train_loss = NaN` (see the **Sanity-run PASS criteria** in `references\u002Ffinetuning.md`)\n- For `evaluate` \u002F `inference`: artifacts under `results_dir`\n\nFor TAO Deploy TensorRT actions (`gen_trt_engine`, TensorRT `evaluate`, and TensorRT `inference`), read `references\u002Ftao-deploy-depth-anything-v2.md` first. Deploy spec templates live in this skill's `references\u002F` folder with the `spec_template_deploy_*.yaml` prefix.\n\n## Training Requirements\n\n- **Valid `dataset_name` values for mono `data_sources`** (case-insensitive): `ThreeDVLM`, `FSD`, `NvCLIP`, `IssacStereo`, `Crestereo`, `Middlebury`, `NYUDV2`, `NYUDV2Relative`, `RelativeMonoDataset`, `MetricMonoDataset`. `NYUDV2` carries metric depth GT (meters) — pair with `MetricDepthAnything`; `NYUDV2Relative` is the same data with relative-depth conventions — pair with `RelativeDepthAnything`.\n- **Monitoring metric:** val\u002Floss\n\n### Per-Action Dataset Requirements\n\n| Action | Spec Key | Source | Files | List? |\n|---|---|---|---|---|\n| evaluate | dataset.test_dataset.data_sources | eval_dataset | data_file: annotations.txt + dataset_name | Yes |\n| inference | dataset.infer_dataset.data_sources | inference_dataset | data_file: annotations.txt + dataset_name | Yes |\n| quantize | dataset.train_dataset.data_sources | train_datasets | data_file: annotations.txt + dataset_name | Yes |\n| quantize | dataset.val_dataset.data_sources | eval_dataset | data_file: annotations.txt + dataset_name | Yes |\n| quantize | dataset.quant_calibration_dataset.images_dir | train_datasets | images.tar.gz | No |\n| train | dataset.train_dataset.data_sources | train_datasets | data_file: annotations.txt + dataset_name | Yes |\n| train | dataset.val_dataset.data_sources | eval_dataset | data_file: annotations.txt + dataset_name | Yes |\n\n### Spec Overrides\n\nData source overrides are **mandatory for every action** — construct the data source paths from the Per-Action Dataset Requirements table above and include them in `spec_overrides`; each `data_sources` entry is a dict with the two mandatory fields `data_file` and `dataset_name`. See `references\u002Fspec-overrides.md` for the full per-action `train` \u002F `evaluate` \u002F `export` \u002F `inference` \u002F `quantize` override blocks and the precision recommendations.\n\n## Eval Dataset\n\nOptional. Val dataset configured via `dataset.val_dataset.data_sources` (each entry needs `data_file` and `dataset_name`).\n\n## Important Parameters\n\nFull parameter glossary (`model.*`, `train.*`, `dataset.*`, `export.*`, `inference.*` fields with options, defaults, and sources) plus the **Pretrained checkpoint loading — use case matrix** live in `references\u002Fparameters.md`. Key starting points: `model.model_type` (default `MetricDepthAnything`), `model.encoder` (default `vitl`), `train.optim.lr` (default 1e-4, AdamW), `train.precision` (`fp32` recommended), `dataset.{train,val,test,infer}_dataset.augmentation.crop_size` (default `[518, 518]`).\n\n## Finetuning Recipes\n\nRelative and Metric variant finetuning recipes — including required spec keys, the metric `dataset.{normalize_depth, min_depth, max_depth}` block required in both train AND export specs, trainer-enforced defaults (`clip_grad_norm: 0.1`, `warmup_steps: 20`, `weight_decay: 1e-4`), sanity-run overrides, and the **Sanity-run PASS criteria** for catching silent `train_loss = NaN` — are in `references\u002Ffinetuning.md`. Both recipes use `train.optim.lr: 5e-6` with `LambdaLR` (the AdamW default `1e-4` is too aggressive when finetuning from a converged\u002Fpretrained backbone).\n\n## Multi-GPU \u002F Multi-Node\n\n**Launch method:** Lightning-managed (single `python` process, Lightning spawns workers).\n\n| Spec Key | Description | Default |\n|----------|-------------|---------|\n| `train.num_gpus` | Number of GPUs | 1 |\n| `train.gpu_ids` | GPU device indices | [0] |\n| `train.num_nodes` | Number of nodes | 1 |\n| `train.distributed_strategy` | `ddp` or `fsdp` | `ddp` |\n\n- `ddp` with activation checkpointing: `find_unused_parameters=False`\n- `ddp` without: `find_unused_parameters=True`\n- `fsdp` forces precision to FP16\n\n**Multi-node env vars** (set by orchestrator): `WORLD_SIZE`, `NODE_RANK`, `MASTER_ADDR`, `MASTER_PORT`, `NUM_GPU_PER_NODE`.\n\n## Export \u002F TRT Defaults\n\n- TRT data types: FP32, BF16 (Ampere SM80+). FP16 is not supported for the ViT-L mono backbone.\n- Recommended TRT precision: `bf16`. Use `fp32` if BF16 hardware is unavailable.\n\nFull TAO Deploy reference: [tao-deploy-depth-anything-v2](references\u002Ftao-deploy-depth-anything-v2.md).\n\n## Hardware\n\nMinimum 1 GPU(s), recommended 2 GPU(s). 24GB+ VRAM per GPU. ViT-Large encoder is memory intensive. Use `fp32` (recommended) or `bf16` (Ampere SM80+, alternative) for training. Activation checkpointing is available for larger inputs.\n\n## Error Patterns\n\nCommon failure signatures and fixes — depth range mismatch, missing pretrained weights, `Key 'encoder' not in 'MonoBackBone'`, missing `dataset_name`, `depth_net_mono: not found`, metric variant hyperparameter sourcing, and the export refuse-to-overwrite ONNX error — are documented in `references\u002Ftroubleshooting.md`.\n\n## Spec Param \u002F Parent Model Inference\n\nModel-specific inference mappings (the full `depth_net_mono.config.json` per-action spec-field → inference-function table, plus `parent_model` \u002F `parent_job_id` resolution guidance) are in `references\u002Fspec-param-inference.md`. These mappings belong in MD, not in `config.json`; generated runners should read that reference and apply the mappings with SDK helpers before `create_job()`.\n",{"name":5,"tags":44,"license":15,"metadata":45,"description":7,"allowed-tools":48,"compatibility":49},[37,21,30],{"author":46,"version":47},"NVIDIA Corporation","0.1.0","Read Bash","Requires docker + nvidia-container-toolkit.","main","7a940b8e5af16c95e5c190a524421d32fc7445a9",".agent\u002Fskills\u002Ftao-train-depth-anything-v2","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui","https:\u002F\u002Fneuroads.com.br","2026-08-12 00:42:44",{"source":8,"skillName":5,"commands":58,"prompt":68,"zipAvailable":69},[59,62,65],{"tool":60,"command":61},"npx","npx skills add headmktlucca-ui\u002Fneuroads-lp --skill tao-train-depth-anything-v2",{"tool":63,"command":64},"bunx","bunx skills add headmktlucca-ui\u002Fneuroads-lp --skill tao-train-depth-anything-v2",{"tool":66,"command":67},"pnpm","pnpm dlx skills add headmktlucca-ui\u002Fneuroads-lp --skill tao-train-depth-anything-v2","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002FSKILL.md",true,[71,77,81,85,89,93,97,101,105,109,113,117,120,124,128,132,136,140,144,148,152,156,160,164,168,172],{"path":72,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":75,"sourceHtmlUrl":76},"BENCHMARK.md","REFERENCE","参考文档",4671,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002FBENCHMARK.md",{"path":78,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":79,"sourceHtmlUrl":80},"evals\u002Fevals.json",793,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Fevals\u002Fevals.json",{"path":82,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":83,"sourceHtmlUrl":84},"references\u002Ffinetuning.md",7265,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Ffinetuning.md",{"path":86,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":87,"sourceHtmlUrl":88},"references\u002Fparameters.md",5396,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Fparameters.md",{"path":90,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":91,"sourceHtmlUrl":92},"references\u002Fskill_info.yaml",1637,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Fskill_info.yaml",{"path":94,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":95,"sourceHtmlUrl":96},"references\u002Fspec_template_deploy.yaml",2020,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Fspec_template_deploy.yaml",{"path":98,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":99,"sourceHtmlUrl":100},"references\u002Fspec_template_evaluate.yaml",5546,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Fspec_template_evaluate.yaml",{"path":102,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":103,"sourceHtmlUrl":104},"references\u002Fspec_template_export.yaml",5615,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Fspec_template_export.yaml",{"path":106,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":107,"sourceHtmlUrl":108},"references\u002Fspec_template_gen_trt_engine.yaml",5663,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Fspec_template_gen_trt_engine.yaml",{"path":110,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":111,"sourceHtmlUrl":112},"references\u002Fspec_template_inference.yaml",5552,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Fspec_template_inference.yaml",{"path":114,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":115,"sourceHtmlUrl":116},"references\u002Fspec_template_quantize.yaml",5381,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Fspec_template_quantize.yaml",{"path":118,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":115,"sourceHtmlUrl":119},"references\u002Fspec_template_train.yaml","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Fspec_template_train.yaml",{"path":121,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":122,"sourceHtmlUrl":123},"references\u002Fspec-overrides.md",3608,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Fspec-overrides.md",{"path":125,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":126,"sourceHtmlUrl":127},"references\u002Fspec-param-inference.md",3234,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Fspec-param-inference.md",{"path":129,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":130,"sourceHtmlUrl":131},"references\u002Ftao-deploy-depth-anything-v2.md",7088,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Ftao-deploy-depth-anything-v2.md",{"path":133,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":134,"sourceHtmlUrl":135},"references\u002Ftao-deploy-depth-anything-v2.skill_info.yaml",2388,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Ftao-deploy-depth-anything-v2.skill_info.yaml",{"path":137,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":138,"sourceHtmlUrl":139},"references\u002Ftroubleshooting.md",1892,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Freferences\u002Ftroubleshooting.md",{"path":141,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":142,"sourceHtmlUrl":143},"schemas\u002Fevaluate.schema.json",106992,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Fschemas\u002Fevaluate.schema.json",{"path":145,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":146,"sourceHtmlUrl":147},"schemas\u002Fexport.schema.json",107656,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Fschemas\u002Fexport.schema.json",{"path":149,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":150,"sourceHtmlUrl":151},"schemas\u002Fgen_trt_engine.schema.json",109557,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Fschemas\u002Fgen_trt_engine.schema.json",{"path":153,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":154,"sourceHtmlUrl":155},"schemas\u002Finference.schema.json",107437,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Fschemas\u002Finference.schema.json",{"path":157,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":158,"sourceHtmlUrl":159},"schemas\u002Fmanifest.json",36696,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Fschemas\u002Fmanifest.json",{"path":161,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":162,"sourceHtmlUrl":163},"schemas\u002Fquantize.schema.json",103802,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Fschemas\u002Fquantize.schema.json",{"path":165,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":166,"sourceHtmlUrl":167},"schemas\u002Ftrain.schema.json",103796,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Fschemas\u002Ftrain.schema.json",{"path":169,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":170,"sourceHtmlUrl":171},"skill-card.md",4056,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Fskill-card.md",{"path":173,"resourceType":174,"resourceTypeDesc":175,"sizeBytes":176,"sourceHtmlUrl":177},"skill.oms.sig","OTHER","其他",10049,"https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.agent\u002Fskills\u002Ftao-train-depth-anything-v2\u002Fskill.oms.sig",[179,206,230,256,274,309],{"skillNo":180,"name":181,"nameZh":6,"description":182,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":183,"licenseFileUrl":6,"sourceHtmlUrl":184,"categories":185,"keywords":194,"publishedAt":39},"01M6PTDCZTXTGSTX4RFK901YSZ","xunit-skill","Generates xUnit.net tests in C#. Covers Fact\u002FTheory, constructor injection, IClassFixture, and FluentAssertions. Use when user mentions \"xUnit\", \"[Fact]\", \"[Theory]\", \"Assert.Equal\", \"C# xUnit\". Triggers on: \"xUnit\", \"[Fact]\", \"[Theory]\", \"Assert.Equal C#\", \"xUnit.net\".","MIT","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fxunit-skill\u002FSKILL.md",[186,191],{"categoryNo":187,"categoryName":188,"taxonomyType":189,"taxonomyTypeDesc":190},"01JBSEEDCAT000000000000008","测试与质量","SCENARIO","使用场景",{"categoryNo":192,"categoryName":193,"taxonomyType":189,"taxonomyTypeDesc":190},"01JBSEEDCAT000000000000023","Skill与智能体开发",[195,196,197,198,199,200,201,202,203,204,205],"xunit","xunit skill","testing","injection","user","use","tests","triggers","unit","unit-testing","C#",{"skillNo":207,"name":208,"nameZh":6,"description":209,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":183,"licenseFileUrl":6,"sourceHtmlUrl":210,"categories":211,"keywords":214,"publishedAt":39},"01M6PTCB59ZTG5FTMWGPYEMMBP","xcuitest-skill","Generates XCUITest UI tests for iOS\u002FiPadOS apps in Swift. Apple's native testing framework for reliable, fast UI automation. Supports local simulators and TestMu AI cloud real devices. Use when user mentions \"XCUITest\", \"XCTest\", \"iOS UI test\", \"Swift test\", \"XCUIApplication\". Triggers on: \"XCUITest\", \"XCTest UI\", \"iOS UI test\", \"Swift UI test\", \"XCUIApplication\", \"TestMu\".","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fxcuitest-skill\u002FSKILL.md",[212,213],{"categoryNo":187,"categoryName":188,"taxonomyType":189,"taxonomyTypeDesc":190},{"categoryNo":192,"categoryName":193,"taxonomyType":189,"taxonomyTypeDesc":190},[215,216,197,217,218,219,220,221,200,222,223,224,225,226,201,202,227,228,33,229],"xcuitest","xcuitest skill","automation","test","testing framework","ios","devices","mobile","mobile-testing","ui test","local","apple","apps","native","objective",{"skillNo":231,"name":232,"nameZh":6,"description":233,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":183,"licenseFileUrl":6,"sourceHtmlUrl":234,"categories":235,"keywords":240,"publishedAt":39},"01M6PTATMNTY9NF4J7KSR76YK6","world-builder","Create and manage generative game zones for Dorothy's Pokemon-style world. Use this skill when creating, updating, or designing game worlds via MCP tools.","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fworld-builder\u002FSKILL.md",[236,239],{"categoryNo":237,"categoryName":238,"taxonomyType":189,"taxonomyTypeDesc":190},"01JBSEEDCAT000000000000015","视觉设计",{"categoryNo":192,"categoryName":193,"taxonomyType":189,"taxonomyTypeDesc":190},[241,242,243,244,245,246,247,248,200,249,250,251,252,253,254,33,255],"builder","world","world builder","MCP","create","mcp tools","Tools","game","pokemon","manage","style","via","theme","zone","generative",{"skillNo":257,"name":258,"nameZh":6,"description":259,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":183,"licenseFileUrl":6,"sourceHtmlUrl":260,"categories":261,"keywords":264,"publishedAt":39},"01M6PT9Z92JTSQEHE1QR0NNW5F","webdriverio-skill","Generates WebdriverIO (WDIO) automation tests in JavaScript or TypeScript. Supports local and TestMu AI cloud. Use when user mentions \"WebdriverIO\", \"WDIO\", \"wdio.conf\", \"browser.url\", \"$\", \"$$\". Triggers on: \"WebdriverIO\", \"WDIO\", \"wdio\", \"browser.$\".","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fwebdriverio-skill\u002FSKILL.md",[262,263],{"categoryNo":187,"categoryName":188,"taxonomyType":189,"taxonomyTypeDesc":190},{"categoryNo":192,"categoryName":193,"taxonomyType":189,"taxonomyTypeDesc":190},[265,197,266,217,267,268,269,270,199,200,271,225,201,202,272,273],"webdriverio skill","typescript","javascript","browser","cloud","e2e","e2e-testing","URL","conf",{"skillNo":275,"name":276,"nameZh":6,"description":277,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":183,"licenseFileUrl":6,"sourceHtmlUrl":278,"categories":279,"keywords":289,"publishedAt":39},"01M6PT9NVK36ZWCQ69YXK3FR4K","web-design-guidelines","Web platform design and accessibility guidelines. Use when building web interfaces, auditing accessibility, implementing responsive layouts, or reviewing web UI code. Triggers on tasks involving HTML, CSS, web components, WCAG compliance, responsive design, or web performance.","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fweb\u002FSKILL.md",[280,283,284],{"categoryNo":281,"categoryName":282,"taxonomyType":189,"taxonomyTypeDesc":190},"01JBSEEDCAT000000000000009","前端开发",{"categoryNo":237,"categoryName":238,"taxonomyType":189,"taxonomyTypeDesc":190},{"categoryNo":285,"categoryName":286,"taxonomyType":287,"taxonomyTypeDesc":288},"01JBSEEDCAT000000000000032","设计师","PROFESSION","适用职业",[290,291,292,293,294,295,296,297,298,299,300,301,302,303,304,305,306,307,202,308],"design","web","guidelines","web design guidelines","web design","css","performance","components","web components","tasks","code","web performance","ui","platform design","html","responsive design","responsive","platform","implementing",{"skillNo":310,"name":311,"nameZh":6,"description":312,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":183,"licenseFileUrl":6,"sourceHtmlUrl":313,"categories":314,"keywords":317,"publishedAt":39},"01M6PT9A3JKRE3X34276F2FW5R","watchos-design-guidelines","Apple Human Interface Guidelines for Apple Watch. Use when building watchOS apps, complications, or workout features. Triggers on tasks involving Watch UI, Digital Crown, glanceable interfaces, or wrist-based interactions.","https:\u002F\u002Fgithub.com\u002Fheadmktlucca-ui\u002Fneuroads-lp\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fwatchos\u002FSKILL.md",[315,316],{"categoryNo":237,"categoryName":238,"taxonomyType":189,"taxonomyTypeDesc":190},{"categoryNo":285,"categoryName":286,"taxonomyType":287,"taxonomyTypeDesc":288},[290,292,318,299,302,200,319,320,226,307,321,202,227,322,33,323,324],"watchos design guidelines","workout","human","watch","interface","platform-design","features"]