[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7AAA8E2BG02TYQMJK37SFCE":3,"skill-same:01M7AAA8E2BG02TYQMJK37SFCE":128},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":33,"commitSha":34,"skillDirectory":35,"forksCount":36,"repositoryHtmlUrl":37,"repositoryHomepageUrl":38,"ownerHtmlUrl":39,"ownerBlogUrl":40,"ownerDescription":41,"updatedAt":42,"install":43,"resources":56},"01M7AAA8E2BG02TYQMJK37SFCE","tao-train-nvdinov2",null,"NVDINOv2 for self-supervised visual representation learning. Trains vision transformers via self-distillation (teacher-student) without labels and produces general-purpose visual features. Use when training, exporting, or running inference for a TAO NVDINOv2 backbone. Trigger phrases include \"train NVDINOv2\", \"self-supervised ViT pretraining\", \"DINOv2 backbone\", \"visual representation learning\".","themusebook\u002Fskills","themusebook","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F349927247?v=4","01M7A5C1RV5S3V5FF6ZACVDNRN","musebook","01M7A5C1SFC3KRZAXNN60XJ7KM",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: tao-train-nvdinov2\ndescription: NVDINOv2 for self-supervised visual representation learning. Trains vision transformers via self-distillation\n  (teacher-student) without labels and produces general-purpose visual features. Use when training, exporting, or\n  running inference for a TAO NVDINOv2 backbone. Trigger phrases include \"train NVDINOv2\", \"self-supervised ViT pretraining\",\n  \"DINOv2 backbone\", \"visual representation learning\".\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- self\n- supervised\n- learning\n---\n\n# NVDINOv2\n\nNVDINOv2 for self-supervised visual representation learning. Trains vision transformers via self-distillation (teacher-student) without labels. Produces general-purpose visual features.\n\nSet train.pretrained_model_path for pretrained ViT weights.\n\nFor TAO Deploy TensorRT actions (`gen_trt_engine`), read `references\u002Ftao-deploy-nvdinov2.md` first. Deploy spec templates live in this skill's `references\u002F` folder with the `spec_template_deploy_*.yaml` prefix.\n\n## Dataclass Schemas\n\nGenerated TAO Core schemas are packaged in `schemas\u002F\u003Caction>.schema.json`, with `schemas\u002Fmanifest.json` listing available actions. Each generated schema also emits `references\u002Fspec_template_\u003Caction>.yaml` from the schema top-level `default` field. AutoML enablement is declared at the model layer in `references\u002Fskill_info.yaml` via `automl_enabled`. Runnable AutoML still requires `schemas\u002Ftrain.schema.json` and `references\u002Fspec_template_train.yaml` to exist and parse. Use the packaged train schema for `automl_default_parameters`, `automl_disabled_parameters`, defaults, min\u002Fmax bounds, enums, option weights, math conditions, dependencies, and popular parameters. Do not expect `~\u002Ftao-core` at runtime; maintainers regenerate schemas\u002Ftemplates before packaging the skill bank.\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. Use `automl_policy: on` by default and only expose `on` \u002F `off` in new launch prompts. Treat phrases like \"turn off AutoML\", \"disable AutoML\", \"no HPO\", or \"plain training\" as `automl_policy: off` for this run only. When `automl_policy: on`, `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 `inference`, `export`, and deploy flows stay in this model skill. The per-run `automl_policy` override does not change model metadata.\n\n## Training Requirements\n\n- **Dataset type:** image_classification\n- **Formats:** ssl\n- **Monitoring metric:** train_loss\n\n### Per-Action Dataset Requirements\n\n| Action | Spec Key | Source | Files | List? |\n|---|---|---|---|---|\n| inference | dataset.test_dataset.images_dir | inference_dataset | images_test.tar.gz | No |\n| train | dataset.train_dataset.images_dir | train_datasets | images_train.tar.gz | No |\n\n### Typical Spec Overrides\n\nData source overrides are **mandatory for train and inference** — the agent MUST construct data source paths from the Per-Action Dataset Requirements table above and include them in `spec_overrides`.\n\n```python\nS3_TRAIN = \"s3:\u002F\u002Fbucket\u002Fdata\u002Ftrain\"\nS3_EVAL = \"s3:\u002F\u002Fbucket\u002Fdata\u002Feval\"\n```\n\n**train (mandatory data sources):**\n```python\n{\n    \"train.num_gpus\": 1,\n    \"train.num_epochs\": 10,\n    \"train.checkpoint_interval\": 10,\n    \"dataset.train_dataset.images_dir\": f\"{S3_TRAIN}\u002Fimages_train.tar.gz\",\n}\n```\n\n**local AutoML validation \u002F smoke run:**\nUse this shape when the goal is to confirm Bayesian launch, metric selection,\nbest-model choice, and checkpoint persistence on local Docker. It keeps the\nrun representative while avoiding the much slower ViT-Large default.\n\n```python\n{\n    \"wandb.enable\": False,\n    \"model.backbone.teacher_type\": \"vit_s\",\n    \"model.backbone.student_type\": \"vit_s\",\n    \"model.backbone.img_size\": 224,\n    \"dataset.batch_size\": 8,\n    \"dataset.workers\": 2,\n    \"train.num_epochs\": 1,\n    \"train.checkpoint_interval\": 1,\n    \"train.num_prototypes\": 1024,\n    \"train.precision\": \"32-true\",\n    \"train.use_custom_attention\": False,\n    \"train.num_gpus\": 1,\n    \"dataset.train_dataset.images_dir\": f\"{S3_TRAIN}\u002Fimages_train.tar.gz\",\n}\n```\n\n**export (mandatory checkpoint handoff):**\n```python\n{\n    \"export.checkpoint\": \"\u003Cselected train\u002FAutoML student_epoch_* checkpoint>\",\n    \"export.onnx_file\": \"\u002Fpath\u002Fto\u002Fresults\u002Fnvdinov2.onnx\",\n}\n```\n\n**inference (mandatory data sources):**\n```python\n{\n    \"inference.checkpoint\": \"\u003Cselected train\u002FAutoML student_epoch_* checkpoint>\",\n    \"model.backbone.teacher_type\": \"\u003Csame value used for train>\",\n    \"model.backbone.student_type\": \"\u003Csame value used for train>\",\n    \"model.backbone.img_size\": \"\u003Csame value used for train>\",\n    \"train.use_custom_attention\": \"\u003Csame value used for train>\",\n    \"dataset.test_dataset.images_dir\": f\"{S3_EVAL}\u002Fimages_test.tar.gz\",\n}\n```\n## Eval Dataset\n\nOptional. SSL training does not use labels. Evaluation is downstream task-specific.\n\n## Important Parameters\n\n- **model.backbone.teacher_type**: Teacher ViT variant. Default vit_l (ViT-Large).\n- **model.backbone.student_type**: Student ViT variant. Default vit_l. Typically matches teacher.\n- **model.backbone.img_size**: Input image size. Default 518. Higher resolution produces better features but costs more memory.\n- **model.backbone.patch_size**: ViT patch size. Default 14.\n- **dataset.batch_size**: Per-GPU batch size. Default 4. SSL training is memory-intensive due to dual (teacher+student) forward passes.\n- **train.layerwise_decay**: Layer-wise learning rate decay. Important for ViT fine-tuning.\n- **train.clip_grad_norm**: Gradient clipping. Important for stable SSL training.\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\n- Strategy: `auto` (Lightning picks best strategy automatically)\n- `sync_batchnorm` is always enabled — critical for SSL training with teacher-student framework\n- Multi-GPU strongly recommended (4-8 GPUs) for meaningful SSL training\n\n**Multi-node env vars** (set by orchestrator): `WORLD_SIZE`, `NODE_RANK`, `MASTER_ADDR`, `MASTER_PORT`, `NUM_GPU_PER_NODE`.\n\n## Hardware\n\nMinimum 4 GPU(s), recommended 8 GPU(s). 40GB+ (A100 recommended) VRAM per GPU. SSL with ViT-Large teacher+student is very memory-intensive. Requires A100 40GB+ GPUs. Multi-GPU strongly recommended.\n\n## Error Patterns\n\n**CUDA out of memory**: ViT-Large teacher+student with img_size=518 requires 40GB+ GPU memory. Reduce batch_size, img_size, or use smaller ViT variant.\n\n**Inference checkpoint has unexpected Lightning keys**: For downstream\n`inference`, pass the selected AutoML run's `student_epoch_*.pth` checkpoint,\nnot `nvdinov2_model_latest.pth`. The latest file is a training checkpoint and\nthe inference loader reports unexpected keys such as `state_dict`, optimizer\nstate, and scheduler state.\n\n**Export checkpoint has unexpected Lightning keys**: Export also consumes the\nselected `student_epoch_*.pth` checkpoint. Use the full `model_epoch_*.pth`\ncheckpoint only for resume\u002Fretrain via `train.resume_training_checkpoint_path`.\n\n**TensorRT engine passed to PyT inference**: The packaged PyT `nvdinov2 inference`\nimplementation only loads `.pth` or `.tlt` model paths. TAO Deploy\n`gen_trt_engine` builds a TensorRT engine for downstream consumers, but the PyT\ninference action does not run on that engine.\n\n**Separate distill action not available**: The current TAO PyT CLI exposes\n`export`, `inference`, `train`, and `default_specs` for NvDINOv2. Do not launch\nor advertise a standalone `nvdinov2 distill` action.\n\n**AutoML metric not found**: TAO's status KPI reports the final training scalar\nas `train_loss`. Use `train_loss` with minimize direction for AutoML selection.\nSome Lightning progress lines also render the same scalar as\n`train_loss_epoch`; treat that as a fallback alias only, not the primary\nmonitoring metric.\n\n**Slow convergence**: SSL needs many epochs. Default 10 is for quick testing; production runs typically use 100+ epochs.\n\n## Spec Param \u002F Parent Model Inference\n\nModel-specific inference mappings belong in this MD file, not in `config.json`. Generated runners should read this section and apply the mappings with SDK helpers before `create_job()`. This mirrors the old microservices `infer_params.py` flow.\n\nModel-specific handoff mappings:\n\n| Action | Spec Field | Inference Function | Meaning |\n|---|---|---|---|\n| export | `encryption_key` | `key` | encryption key |\n| export | `export.checkpoint` | `parent_model` | selected `student_epoch_*.pth` checkpoint from the parent job results folder |\n| export | `export.onnx_file` | `create_onnx_file` | output ONNX path |\n| export | `results_dir` | `output_dir` | current job results directory |\n| inference | `encryption_key` | `key` | encryption key |\n| inference | `inference.checkpoint` | `parent_model` | selected `student_epoch_*.pth` checkpoint from the parent job results folder |\n| inference | `results_dir` | `output_dir` | current job results directory |\n| train | `encryption_key` | `key` | encryption key |\n| train | `results_dir` | `output_dir` | current job results directory |\n| train | `train.pretrained_model_path` | `ptm_if_no_resume_model` | PTM when no resume checkpoint exists |\n| train | `train.resume_training_checkpoint_path` | `resume_model` | selected full `model_epoch_*.pth` training checkpoint from the current job results folder |\n\nFor `parent_model` or `parent_model_folder`, pass the upstream train\u002Fexport\u002FAutoML child job id as `parent_job_id`. The SDK lists the parent result folder, filters checkpoint artifacts, and returns the selected model file or folder. Do not add these mappings back to `config.json` and do not patch generated runner scripts to guess checkpoint paths.\n\n## Deployment\n\n- [tao-deploy-nvdinov2](references\u002Ftao-deploy-nvdinov2.md)\n","# NVDINOv2\n\nNVDINOv2 for self-supervised visual representation learning. Trains vision transformers via self-distillation (teacher-student) without labels. Produces general-purpose visual features.\n\nSet train.pretrained_model_path for pretrained ViT weights.\n\nFor TAO Deploy TensorRT actions (`gen_trt_engine`), read `references\u002Ftao-deploy-nvdinov2.md` first. Deploy spec templates live in this skill's `references\u002F` folder with the `spec_template_deploy_*.yaml` prefix.\n\n## Dataclass Schemas\n\nGenerated TAO Core schemas are packaged in `schemas\u002F\u003Caction>.schema.json`, with `schemas\u002Fmanifest.json` listing available actions. Each generated schema also emits `references\u002Fspec_template_\u003Caction>.yaml` from the schema top-level `default` field. AutoML enablement is declared at the model layer in `references\u002Fskill_info.yaml` via `automl_enabled`. Runnable AutoML still requires `schemas\u002Ftrain.schema.json` and `references\u002Fspec_template_train.yaml` to exist and parse. Use the packaged train schema for `automl_default_parameters`, `automl_disabled_parameters`, defaults, min\u002Fmax bounds, enums, option weights, math conditions, dependencies, and popular parameters. Do not expect `~\u002Ftao-core` at runtime; maintainers regenerate schemas\u002Ftemplates before packaging the skill bank.\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. Use `automl_policy: on` by default and only expose `on` \u002F `off` in new launch prompts. Treat phrases like \"turn off AutoML\", \"disable AutoML\", \"no HPO\", or \"plain training\" as `automl_policy: off` for this run only. When `automl_policy: on`, `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 `inference`, `export`, and deploy flows stay in this model skill. The per-run `automl_policy` override does not change model metadata.\n\n## Training Requirements\n\n- **Dataset type:** image_classification\n- **Formats:** ssl\n- **Monitoring metric:** train_loss\n\n### Per-Action Dataset Requirements\n\n| Action | Spec Key | Source | Files | List? |\n|---|---|---|---|---|\n| inference | dataset.test_dataset.images_dir | inference_dataset | images_test.tar.gz | No |\n| train | dataset.train_dataset.images_dir | train_datasets | images_train.tar.gz | No |\n\n### Typical Spec Overrides\n\nData source overrides are **mandatory for train and inference** — the agent MUST construct data source paths from the Per-Action Dataset Requirements table above and include them in `spec_overrides`.\n\n```python\nS3_TRAIN = \"s3:\u002F\u002Fbucket\u002Fdata\u002Ftrain\"\nS3_EVAL = \"s3:\u002F\u002Fbucket\u002Fdata\u002Feval\"\n```\n\n**train (mandatory data sources):**\n```python\n{\n    \"train.num_gpus\": 1,\n    \"train.num_epochs\": 10,\n    \"train.checkpoint_interval\": 10,\n    \"dataset.train_dataset.images_dir\": f\"{S3_TRAIN}\u002Fimages_train.tar.gz\",\n}\n```\n\n**local AutoML validation \u002F smoke run:**\nUse this shape when the goal is to confirm Bayesian launch, metric selection,\nbest-model choice, and checkpoint persistence on local Docker. It keeps the\nrun representative while avoiding the much slower ViT-Large default.\n\n```python\n{\n    \"wandb.enable\": False,\n    \"model.backbone.teacher_type\": \"vit_s\",\n    \"model.backbone.student_type\": \"vit_s\",\n    \"model.backbone.img_size\": 224,\n    \"dataset.batch_size\": 8,\n    \"dataset.workers\": 2,\n    \"train.num_epochs\": 1,\n    \"train.checkpoint_interval\": 1,\n    \"train.num_prototypes\": 1024,\n    \"train.precision\": \"32-true\",\n    \"train.use_custom_attention\": False,\n    \"train.num_gpus\": 1,\n    \"dataset.train_dataset.images_dir\": f\"{S3_TRAIN}\u002Fimages_train.tar.gz\",\n}\n```\n\n**export (mandatory checkpoint handoff):**\n```python\n{\n    \"export.checkpoint\": \"\u003Cselected train\u002FAutoML student_epoch_* checkpoint>\",\n    \"export.onnx_file\": \"\u002Fpath\u002Fto\u002Fresults\u002Fnvdinov2.onnx\",\n}\n```\n\n**inference (mandatory data sources):**\n```python\n{\n    \"inference.checkpoint\": \"\u003Cselected train\u002FAutoML student_epoch_* checkpoint>\",\n    \"model.backbone.teacher_type\": \"\u003Csame value used for train>\",\n    \"model.backbone.student_type\": \"\u003Csame value used for train>\",\n    \"model.backbone.img_size\": \"\u003Csame value used for train>\",\n    \"train.use_custom_attention\": \"\u003Csame value used for train>\",\n    \"dataset.test_dataset.images_dir\": f\"{S3_EVAL}\u002Fimages_test.tar.gz\",\n}\n```\n## Eval Dataset\n\nOptional. SSL training does not use labels. Evaluation is downstream task-specific.\n\n## Important Parameters\n\n- **model.backbone.teacher_type**: Teacher ViT variant. Default vit_l (ViT-Large).\n- **model.backbone.student_type**: Student ViT variant. Default vit_l. Typically matches teacher.\n- **model.backbone.img_size**: Input image size. Default 518. Higher resolution produces better features but costs more memory.\n- **model.backbone.patch_size**: ViT patch size. Default 14.\n- **dataset.batch_size**: Per-GPU batch size. Default 4. SSL training is memory-intensive due to dual (teacher+student) forward passes.\n- **train.layerwise_decay**: Layer-wise learning rate decay. Important for ViT fine-tuning.\n- **train.clip_grad_norm**: Gradient clipping. Important for stable SSL training.\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\n- Strategy: `auto` (Lightning picks best strategy automatically)\n- `sync_batchnorm` is always enabled — critical for SSL training with teacher-student framework\n- Multi-GPU strongly recommended (4-8 GPUs) for meaningful SSL training\n\n**Multi-node env vars** (set by orchestrator): `WORLD_SIZE`, `NODE_RANK`, `MASTER_ADDR`, `MASTER_PORT`, `NUM_GPU_PER_NODE`.\n\n## Hardware\n\nMinimum 4 GPU(s), recommended 8 GPU(s). 40GB+ (A100 recommended) VRAM per GPU. SSL with ViT-Large teacher+student is very memory-intensive. Requires A100 40GB+ GPUs. Multi-GPU strongly recommended.\n\n## Error Patterns\n\n**CUDA out of memory**: ViT-Large teacher+student with img_size=518 requires 40GB+ GPU memory. Reduce batch_size, img_size, or use smaller ViT variant.\n\n**Inference checkpoint has unexpected Lightning keys**: For downstream\n`inference`, pass the selected AutoML run's `student_epoch_*.pth` checkpoint,\nnot `nvdinov2_model_latest.pth`. The latest file is a training checkpoint and\nthe inference loader reports unexpected keys such as `state_dict`, optimizer\nstate, and scheduler state.\n\n**Export checkpoint has unexpected Lightning keys**: Export also consumes the\nselected `student_epoch_*.pth` checkpoint. Use the full `model_epoch_*.pth`\ncheckpoint only for resume\u002Fretrain via `train.resume_training_checkpoint_path`.\n\n**TensorRT engine passed to PyT inference**: The packaged PyT `nvdinov2 inference`\nimplementation only loads `.pth` or `.tlt` model paths. TAO Deploy\n`gen_trt_engine` builds a TensorRT engine for downstream consumers, but the PyT\ninference action does not run on that engine.\n\n**Separate distill action not available**: The current TAO PyT CLI exposes\n`export`, `inference`, `train`, and `default_specs` for NvDINOv2. Do not launch\nor advertise a standalone `nvdinov2 distill` action.\n\n**AutoML metric not found**: TAO's status KPI reports the final training scalar\nas `train_loss`. Use `train_loss` with minimize direction for AutoML selection.\nSome Lightning progress lines also render the same scalar as\n`train_loss_epoch`; treat that as a fallback alias only, not the primary\nmonitoring metric.\n\n**Slow convergence**: SSL needs many epochs. Default 10 is for quick testing; production runs typically use 100+ epochs.\n\n## Spec Param \u002F Parent Model Inference\n\nModel-specific inference mappings belong in this MD file, not in `config.json`. Generated runners should read this section and apply the mappings with SDK helpers before `create_job()`. This mirrors the old microservices `infer_params.py` flow.\n\nModel-specific handoff mappings:\n\n| Action | Spec Field | Inference Function | Meaning |\n|---|---|---|---|\n| export | `encryption_key` | `key` | encryption key |\n| export | `export.checkpoint` | `parent_model` | selected `student_epoch_*.pth` checkpoint from the parent job results folder |\n| export | `export.onnx_file` | `create_onnx_file` | output ONNX path |\n| export | `results_dir` | `output_dir` | current job results directory |\n| inference | `encryption_key` | `key` | encryption key |\n| inference | `inference.checkpoint` | `parent_model` | selected `student_epoch_*.pth` checkpoint from the parent job results folder |\n| inference | `results_dir` | `output_dir` | current job results directory |\n| train | `encryption_key` | `key` | encryption key |\n| train | `results_dir` | `output_dir` | current job results directory |\n| train | `train.pretrained_model_path` | `ptm_if_no_resume_model` | PTM when no resume checkpoint exists |\n| train | `train.resume_training_checkpoint_path` | `resume_model` | selected full `model_epoch_*.pth` training checkpoint from the current job results folder |\n\nFor `parent_model` or `parent_model_folder`, pass the upstream train\u002Fexport\u002FAutoML child job id as `parent_job_id`. The SDK lists the parent result folder, filters checkpoint artifacts, and returns the selected model file or folder. Do not add these mappings back to `config.json` and do not patch generated runner scripts to guess checkpoint paths.\n\n## Deployment\n\n- [tao-deploy-nvdinov2](references\u002Ftao-deploy-nvdinov2.md)\n",{"name":5,"tags":24,"license":15,"metadata":28,"description":7,"allowed-tools":31,"compatibility":32},[25,26,27],"self","supervised","learning",{"author":29,"version":30},"NVIDIA Corporation","0.1.0","Read Bash","Requires docker + nvidia-container-toolkit.","main","1c523aba5cdf867a79f078847075492d7778092a","skills\u002Fnvidia\u002Ftao-train-nvdinov2",1,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills","https:\u002F\u002Fskills.x402.wtf","https:\u002F\u002Fgithub.com\u002Fthemusebook","https:\u002F\u002Fmusebook.trade","The official Github of musebook","2026-11-02 23:42:26",{"source":8,"skillName":5,"commands":44,"prompt":54,"zipAvailable":55},[45,48,51],{"tool":46,"command":47},"npx","npx skills add themusebook\u002Fskills --skill tao-train-nvdinov2",{"tool":49,"command":50},"bunx","bunx skills add themusebook\u002Fskills --skill tao-train-nvdinov2",{"tool":52,"command":53},"pnpm","pnpm dlx skills add themusebook\u002Fskills --skill tao-train-nvdinov2","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002FSKILL.md",true,[57,63,67,71,75,79,83,87,90,94,98,102,106,110,114,118,122],{"path":58,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":61,"sourceHtmlUrl":62},"BENCHMARK.md","REFERENCE","参考文档",4064,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002FBENCHMARK.md",{"path":64,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":65,"sourceHtmlUrl":66},"evals\u002Fevals.json",796,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Fevals\u002Fevals.json",{"path":68,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":69,"sourceHtmlUrl":70},"references\u002Fskill_info.yaml",2734,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Freferences\u002Fskill_info.yaml",{"path":72,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":73,"sourceHtmlUrl":74},"references\u002Fspec_template_deploy_gen_trt_engine.yaml",514,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Freferences\u002Fspec_template_deploy_gen_trt_engine.yaml",{"path":76,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":77,"sourceHtmlUrl":78},"references\u002Fspec_template_distill.yaml",2113,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Freferences\u002Fspec_template_distill.yaml",{"path":80,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":81,"sourceHtmlUrl":82},"references\u002Fspec_template_export.yaml",2314,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Freferences\u002Fspec_template_export.yaml",{"path":84,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":85,"sourceHtmlUrl":86},"references\u002Fspec_template_inference.yaml",2264,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Freferences\u002Fspec_template_inference.yaml",{"path":88,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":77,"sourceHtmlUrl":89},"references\u002Fspec_template_train.yaml","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Freferences\u002Fspec_template_train.yaml",{"path":91,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":92,"sourceHtmlUrl":93},"references\u002Ftao-deploy-nvdinov2.md",3852,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Freferences\u002Ftao-deploy-nvdinov2.md",{"path":95,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":96,"sourceHtmlUrl":97},"references\u002Ftao-deploy-nvdinov2.skill_info.yaml",1605,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Freferences\u002Ftao-deploy-nvdinov2.skill_info.yaml",{"path":99,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":100,"sourceHtmlUrl":101},"schemas\u002Fdistill.schema.json",40774,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Fschemas\u002Fdistill.schema.json",{"path":103,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":104,"sourceHtmlUrl":105},"schemas\u002Fexport.schema.json",43536,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Fschemas\u002Fexport.schema.json",{"path":107,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":108,"sourceHtmlUrl":109},"schemas\u002Finference.schema.json",43698,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Fschemas\u002Finference.schema.json",{"path":111,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":112,"sourceHtmlUrl":113},"schemas\u002Fmanifest.json",12582,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Fschemas\u002Fmanifest.json",{"path":115,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":116,"sourceHtmlUrl":117},"schemas\u002Ftrain.schema.json",40770,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Fschemas\u002Ftrain.schema.json",{"path":119,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":120,"sourceHtmlUrl":121},"skill-card.md",3819,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Fskill-card.md",{"path":123,"resourceType":124,"resourceTypeDesc":125,"sizeBytes":126,"sourceHtmlUrl":127},"skill.oms.sig","OTHER","其他",7813,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Ftao-train-nvdinov2\u002Fskill.oms.sig",[129,138,145,152,159,166],{"skillNo":130,"name":131,"nameZh":6,"description":132,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":133,"licenseFileUrl":134,"sourceHtmlUrl":135,"categories":136,"keywords":137,"publishedAt":19},"01M7AB34SPKVSDRRG009PM1CZH","youtube-clipper","YouTube 视频智能剪辑工具。下载视频和字幕，AI 分析生成精细章节（几分钟级别）， 用户选择片段后自动剪辑、翻译字幕为中英双语、烧录字幕到视频，并生成总结文案。 使用场景：当用户需要剪辑 YouTube 视频、生成短视频片段、制作双语字幕版本时。 关键词：视频剪辑、YouTube、字幕翻译、双语字幕、视频下载、clip video","MIT","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fyoutube-clipper\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fyoutube-clipper\u002FSKILL.md",[],[],{"skillNo":139,"name":140,"nameZh":6,"description":141,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":133,"licenseFileUrl":6,"sourceHtmlUrl":142,"categories":143,"keywords":144,"publishedAt":19},"01M7AB18BH6N7KR7939HRNWVMQ","viem-integration","Integrate EVM blockchains using viem. Use when user says \"read blockchain data\", \"send transaction\", \"interact with smart contract\", \"connect to Ethereum\", \"use viem\", \"use wagmi\", \"wallet integration\", \"viem setup\", or mentions blockchain\u002FEVM development with TypeScript.","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fviem-integration\u002FSKILL.md",[],[],{"skillNo":146,"name":147,"nameZh":6,"description":148,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":133,"licenseFileUrl":6,"sourceHtmlUrl":149,"categories":150,"keywords":151,"publishedAt":19},"01M7AB12GHGP48WCNXDQQ9222H","v4-security-foundations","Security-first Uniswap v4 hook development. Use when user mentions \"v4 hooks\", \"hook security\", \"PoolManager\", \"beforeSwap\", \"afterSwap\", or asks about V4 hook best practices, vulnerabilities, or audit requirements.","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fv4-security-foundations\u002FSKILL.md",[],[],{"skillNo":153,"name":154,"nameZh":6,"description":155,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":133,"licenseFileUrl":6,"sourceHtmlUrl":156,"categories":157,"keywords":158,"publishedAt":19},"01M7AB0XM3PSR7V3Q2ZRR0Q0FB","v4-sdk-integration","App-layer SDK guide for building swap and liquidity experiences directly with the Uniswap v4 SDK. Use when user asks about \"v4 sdk\", \"uniswap v4\", \"v4 swap\", \"v4 liquidity\", \"PoolManager\", \"V4Planner\", \"StateView\", \"PositionManager\", \"pool state\", \"v4 position\", \"uniswap sdk\", or when building swap\u002Fliquidity UX directly with SDKs rather than via the Trading API.","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fv4-sdk-integration\u002FSKILL.md",[],[],{"skillNo":160,"name":161,"nameZh":6,"description":162,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":133,"licenseFileUrl":6,"sourceHtmlUrl":163,"categories":164,"keywords":165,"publishedAt":19},"01M7AB0SGA93ASA588DJW4P9WT","v4-hook-generator","Generate Uniswap v4 hook contracts via OpenZeppelin MCP. Use when building custom swap logic, async swaps, hook-owned liquidity, custom curves, dynamic fees, MEV protection, limit orders, or oracle hooks.","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fv4-hook-generator\u002FSKILL.md",[],[],{"skillNo":167,"name":168,"nameZh":6,"description":169,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":133,"licenseFileUrl":6,"sourceHtmlUrl":170,"categories":171,"keywords":172,"publishedAt":19},"01M7AAZYH75H4CS03WSW2EZ8DM","testing","For testing with Light Protocol programs and clients on localnet, devnet, and mainnet validation.","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Ftesting\u002FSKILL.md",[],[]]