[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7CQR25MXGENPD8BEQQHM9MN":3,"skill-same:01M7CQR25MXGENPD8BEQQHM9MN":61},{"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":6,"ownerHtmlUrl":38,"ownerBlogUrl":6,"ownerDescription":39,"updatedAt":40,"install":41,"resources":54},"01M7CQR25MXGENPD8BEQQHM9MN","tao-train-mask-auto-encoder",null,"Masked Auto-Encoder (MAE) for self-supervised pretraining and fine-tuning. Masks random patches and reconstructs them to learn visual representations; supports pretrain and finetune stages. Use when training, evaluating, exporting, or running inference for a TAO MAE backbone. Trigger phrases include \"pretrain MAE\", \"self-supervised vision pretraining\", \"Masked Autoencoder\", \"Mask Auto-Encoder\", \"MAE fine-tune\".","om-scogo\u002Fskillsh-scraper","om-scogo","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F160466840?v=4","01M79AWB6DFVCQ7RK70CS34B96","Om Achrekar","01M79AWB6ZM9SKR8KDH1WSSM22",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fnvidia-skills\u002Ftao-train-mask-auto-encoder\u002FSkill.md",[],[],"2026-11-15 14:19:17",false,"---\nname: tao-train-mask-auto-encoder\ndescription: Masked Auto-Encoder (MAE) for self-supervised pretraining and fine-tuning. Masks random patches and reconstructs\n  them to learn visual representations; supports pretrain and finetune stages. Use when training, evaluating, exporting, or\n  running inference for a TAO MAE backbone. Trigger phrases include \"pretrain MAE\", \"self-supervised vision pretraining\",\n  \"Masked Autoencoder\", \"Mask Auto-Encoder\", \"MAE fine-tune\".\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# MAE\n\nMAE (Masked Autoencoder) for self-supervised pretraining and fine-tuning. Masks random patches and reconstructs them to learn visual representations. Supports pretrain and finetune stages.\n\nSet train.pretrained_model_path for pretrained MAE weights when fine-tuning.\n\nFor TAO Deploy TensorRT actions (`gen_trt_engine`), read `references\u002Ftao-deploy-mask-auto-encoder.md` first. Deploy spec templates live in this skill's `references\u002F` folder with the `spec_template_deploy_*.yaml` prefix.\n\nThe parent PyTorch `mae` CLI supports `train`, `evaluate`, `inference`, and\n`export`. Build TensorRT engines through the deploy workflow, not the model skill.\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 `evaluate`, `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- **Accepted dataset intents:** training, evaluation, testing\n- **Monitoring metric:** train_loss\n\n### Per-Action Dataset Requirements\n\n| Action | Spec Key | Source | Files | List? |\n|---|---|---|---|---|\n| train | dataset.train_data_sources | train_datasets | images_train.tar.gz | No |\n| train | dataset.val_data_sources | eval_dataset | images_val.tar.gz | No |\n| evaluate | dataset.val_data_sources | eval_dataset | images_val.tar.gz | No |\n| inference | dataset.test_data_sources | inference_dataset | images_test.tar.gz | No |\n\nFor SDK\u002Fapp job inputs, the `images_*.tar.gz` archives are uploaded as the\naction inputs. For direct local Docker runs against host-mounted data, extract\nthe archives first and point `dataset.train_data_sources`,\n`dataset.val_data_sources`, and `dataset.test_data_sources` at the extracted\n`images_train`, `images_val`, and `images_test` folders. Passing a local tar\npath directly to the MAE CLI can produce a zero-sample dataloader because the\nlocal dataloader does not unpack that archive path.\n\n### Typical Spec Overrides\n\nData source overrides are **mandatory for every action** — 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    \"dataset.train_data_sources\": f\"{S3_TRAIN}\u002Fimages_train.tar.gz\",\n    \"dataset.val_data_sources\": f\"{S3_EVAL}\u002Fimages_val.tar.gz\",\n    \"train.num_epochs\": 10,\n    \"train.optim.lr\": 2e-4,\n}\n```\n\n**evaluate (mandatory data sources):**\n```python\n{\n    \"dataset.val_data_sources\": f\"{S3_EVAL}\u002Fimages_val.tar.gz\",\n    \"evaluate.checkpoint\": \"\u003Cselected train\u002FAutoML checkpoint>\",\n    \"train.stage\": \"finetune\",\n}\n```\n\n**inference (mandatory data sources):**\n```python\n{\n    \"dataset.test_data_sources\": f\"{S3_EVAL}\u002Fimages_test.tar.gz\",\n    \"inference.checkpoint\": \"\u003Cselected train\u002FAutoML checkpoint>\",\n    \"train.stage\": \"finetune\",\n}\n```\n\n## Eval Dataset\n\nOptional. Pretraining does not need eval data. Fine-tuning optionally uses val set.\n\n## Important Parameters\n\n- **train.stage**: Training stage. Options: pretrain, finetune. Pretrain learns representations via masking. Finetune adds a classification head.\n- **model.arch**: Architecture. Default convnextv2_base. For local smoke\n  AutoML, use `convnextv2_atto` rather than unsupported names such as\n  `vit_tiny_patch16`. Supported families include `vit_base_patch16` and larger\n  ViTs, ConvNeXtV2 atto\u002Ffemto\u002Fpico\u002Fnano\u002Ftiny\u002Fbase\u002Flarge\u002Fhuge, and Hiera\n  tiny\u002Fsmall\u002Fbase\u002Flarge\u002Fhuge.\n- **model.num_classes**: Number of classes for fine-tuning. Default 1000 (ImageNet). Only relevant in finetune stage.\n- **model.mask_ratio**: Fraction of patches to mask during pretraining. Typically 0.75.\n- **model.norm_pix_loss**: Whether to normalize pixel values in reconstruction loss.\n- **dataset.augmentation.input_size**: Keep the local smoke profile at 224\n  for ConvNeXtV2 MAE. Reducing to 112 can make the MAE mask grid incompatible\n  with feature-map dimensions.\n- MAE does not expose a `dataset.workers` spec field. Do not add it to\n  smoke-test overrides; Hydra rejects unknown dataset keys before training.\n- **train.optim.lr**: Learning rate. Default 2e-4.\n- **dataset.augmentation**: Augmentation settings including mixup, cutmix for fine-tuning.\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` uses `find_unused_parameters=True`\n- `fsdp` forces FP16\n- Multi-GPU strongly recommended for pretraining (large batch sizes needed)\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 2 GPU(s), recommended 8 GPU(s). 24GB+ (A100 recommended) VRAM per GPU. MAE pretraining benefits from large batch sizes across many GPUs. Fine-tuning is more modest in resource requirements.\n\n## Error Patterns\n\n**Stage mismatch**: Ensure train.stage matches your intent (pretrain vs finetune). Fine-tuning without a pretrained_model_path trains from scratch.\n\n**Inference with pretrain checkpoints**: The MAE predict dataloader raises\n`NotImplementedError` for `train.stage: pretrain`. Use a `finetune` checkpoint\nfor inference and classification-style evaluation, or restrict a pretrain-only\nrun to train\u002Fevaluate\u002Fexport.\n\n**num_classes mismatch (finetune only)**: Ensure model.num_classes matches your dataset class count when fine-tuning.\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\nInference mappings from TAO Core `mae.config.json`:\n\n| Action | Spec Field | Inference Function | Meaning |\n|---|---|---|---|\n| evaluate | `encryption_key` | `key` | encryption key |\n| evaluate | `evaluate.checkpoint` | `parent_model` | model file inferred from the parent job results folder |\n| evaluate | `evaluate.trt_engine` | `parent_model` | model file inferred from the parent job results folder |\n| evaluate | `results_dir` | `output_dir` | current job results directory |\n| export | `encryption_key` | `key` | encryption key |\n| export | `export.checkpoint` | `parent_model` | model file inferred 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` | model file inferred from the parent job results folder |\n| inference | `inference.trt_engine` | `parent_model` | model file inferred 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` | model file inferred 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\nWhen resolving checkpoints outside the SDK resolver, select the intended\nepoch\u002Fstep artifact exactly, for example `model_epoch_000_step_00099.pth`.\nUse the `convnextv2_atto_latest.pth` or other latest symlink only when latest\nis explicitly requested. Carry `train.stage`, `model.arch`, `model.num_classes`,\nand export input size forward into evaluate, inference, export, and deploy\nspecs so the checkpoint and ONNX\u002Fengine shapes match.\n\n## Deployment\n\n- [tao-deploy-mask-auto-encoder](references\u002Ftao-deploy-mask-auto-encoder.md)\n","# MAE\n\nMAE (Masked Autoencoder) for self-supervised pretraining and fine-tuning. Masks random patches and reconstructs them to learn visual representations. Supports pretrain and finetune stages.\n\nSet train.pretrained_model_path for pretrained MAE weights when fine-tuning.\n\nFor TAO Deploy TensorRT actions (`gen_trt_engine`), read `references\u002Ftao-deploy-mask-auto-encoder.md` first. Deploy spec templates live in this skill's `references\u002F` folder with the `spec_template_deploy_*.yaml` prefix.\n\nThe parent PyTorch `mae` CLI supports `train`, `evaluate`, `inference`, and\n`export`. Build TensorRT engines through the deploy workflow, not the model skill.\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 `evaluate`, `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- **Accepted dataset intents:** training, evaluation, testing\n- **Monitoring metric:** train_loss\n\n### Per-Action Dataset Requirements\n\n| Action | Spec Key | Source | Files | List? |\n|---|---|---|---|---|\n| train | dataset.train_data_sources | train_datasets | images_train.tar.gz | No |\n| train | dataset.val_data_sources | eval_dataset | images_val.tar.gz | No |\n| evaluate | dataset.val_data_sources | eval_dataset | images_val.tar.gz | No |\n| inference | dataset.test_data_sources | inference_dataset | images_test.tar.gz | No |\n\nFor SDK\u002Fapp job inputs, the `images_*.tar.gz` archives are uploaded as the\naction inputs. For direct local Docker runs against host-mounted data, extract\nthe archives first and point `dataset.train_data_sources`,\n`dataset.val_data_sources`, and `dataset.test_data_sources` at the extracted\n`images_train`, `images_val`, and `images_test` folders. Passing a local tar\npath directly to the MAE CLI can produce a zero-sample dataloader because the\nlocal dataloader does not unpack that archive path.\n\n### Typical Spec Overrides\n\nData source overrides are **mandatory for every action** — 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    \"dataset.train_data_sources\": f\"{S3_TRAIN}\u002Fimages_train.tar.gz\",\n    \"dataset.val_data_sources\": f\"{S3_EVAL}\u002Fimages_val.tar.gz\",\n    \"train.num_epochs\": 10,\n    \"train.optim.lr\": 2e-4,\n}\n```\n\n**evaluate (mandatory data sources):**\n```python\n{\n    \"dataset.val_data_sources\": f\"{S3_EVAL}\u002Fimages_val.tar.gz\",\n    \"evaluate.checkpoint\": \"\u003Cselected train\u002FAutoML checkpoint>\",\n    \"train.stage\": \"finetune\",\n}\n```\n\n**inference (mandatory data sources):**\n```python\n{\n    \"dataset.test_data_sources\": f\"{S3_EVAL}\u002Fimages_test.tar.gz\",\n    \"inference.checkpoint\": \"\u003Cselected train\u002FAutoML checkpoint>\",\n    \"train.stage\": \"finetune\",\n}\n```\n\n## Eval Dataset\n\nOptional. Pretraining does not need eval data. Fine-tuning optionally uses val set.\n\n## Important Parameters\n\n- **train.stage**: Training stage. Options: pretrain, finetune. Pretrain learns representations via masking. Finetune adds a classification head.\n- **model.arch**: Architecture. Default convnextv2_base. For local smoke\n  AutoML, use `convnextv2_atto` rather than unsupported names such as\n  `vit_tiny_patch16`. Supported families include `vit_base_patch16` and larger\n  ViTs, ConvNeXtV2 atto\u002Ffemto\u002Fpico\u002Fnano\u002Ftiny\u002Fbase\u002Flarge\u002Fhuge, and Hiera\n  tiny\u002Fsmall\u002Fbase\u002Flarge\u002Fhuge.\n- **model.num_classes**: Number of classes for fine-tuning. Default 1000 (ImageNet). Only relevant in finetune stage.\n- **model.mask_ratio**: Fraction of patches to mask during pretraining. Typically 0.75.\n- **model.norm_pix_loss**: Whether to normalize pixel values in reconstruction loss.\n- **dataset.augmentation.input_size**: Keep the local smoke profile at 224\n  for ConvNeXtV2 MAE. Reducing to 112 can make the MAE mask grid incompatible\n  with feature-map dimensions.\n- MAE does not expose a `dataset.workers` spec field. Do not add it to\n  smoke-test overrides; Hydra rejects unknown dataset keys before training.\n- **train.optim.lr**: Learning rate. Default 2e-4.\n- **dataset.augmentation**: Augmentation settings including mixup, cutmix for fine-tuning.\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` uses `find_unused_parameters=True`\n- `fsdp` forces FP16\n- Multi-GPU strongly recommended for pretraining (large batch sizes needed)\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 2 GPU(s), recommended 8 GPU(s). 24GB+ (A100 recommended) VRAM per GPU. MAE pretraining benefits from large batch sizes across many GPUs. Fine-tuning is more modest in resource requirements.\n\n## Error Patterns\n\n**Stage mismatch**: Ensure train.stage matches your intent (pretrain vs finetune). Fine-tuning without a pretrained_model_path trains from scratch.\n\n**Inference with pretrain checkpoints**: The MAE predict dataloader raises\n`NotImplementedError` for `train.stage: pretrain`. Use a `finetune` checkpoint\nfor inference and classification-style evaluation, or restrict a pretrain-only\nrun to train\u002Fevaluate\u002Fexport.\n\n**num_classes mismatch (finetune only)**: Ensure model.num_classes matches your dataset class count when fine-tuning.\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\nInference mappings from TAO Core `mae.config.json`:\n\n| Action | Spec Field | Inference Function | Meaning |\n|---|---|---|---|\n| evaluate | `encryption_key` | `key` | encryption key |\n| evaluate | `evaluate.checkpoint` | `parent_model` | model file inferred from the parent job results folder |\n| evaluate | `evaluate.trt_engine` | `parent_model` | model file inferred from the parent job results folder |\n| evaluate | `results_dir` | `output_dir` | current job results directory |\n| export | `encryption_key` | `key` | encryption key |\n| export | `export.checkpoint` | `parent_model` | model file inferred 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` | model file inferred from the parent job results folder |\n| inference | `inference.trt_engine` | `parent_model` | model file inferred 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` | model file inferred 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\nWhen resolving checkpoints outside the SDK resolver, select the intended\nepoch\u002Fstep artifact exactly, for example `model_epoch_000_step_00099.pth`.\nUse the `convnextv2_atto_latest.pth` or other latest symlink only when latest\nis explicitly requested. Carry `train.stage`, `model.arch`, `model.num_classes`,\nand export input size forward into evaluate, inference, export, and deploy\nspecs so the checkpoint and ONNX\u002Fengine shapes match.\n\n## Deployment\n\n- [tao-deploy-mask-auto-encoder](references\u002Ftao-deploy-mask-auto-encoder.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","8859218955d9504cd18b59e7892100718039d3b4","data\u002Fskills\u002Fnvidia-skills\u002Ftao-train-mask-auto-encoder",1,"https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper","https:\u002F\u002Fgithub.com\u002Fom-scogo","SDE-1 @ Scogo ","2026-07-18 12:22:46",{"source":8,"skillName":5,"commands":42,"prompt":52,"zipAvailable":53},[43,46,49],{"tool":44,"command":45},"npx","npx skills add om-scogo\u002Fskillsh-scraper --skill tao-train-mask-auto-encoder",{"tool":47,"command":48},"bunx","bunx skills add om-scogo\u002Fskillsh-scraper --skill tao-train-mask-auto-encoder",{"tool":50,"command":51},"pnpm","pnpm dlx skills add om-scogo\u002Fskillsh-scraper --skill tao-train-mask-auto-encoder","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fnvidia-skills\u002Ftao-train-mask-auto-encoder\u002FSkill.md",true,[55],{"path":56,"resourceType":57,"resourceTypeDesc":58,"sizeBytes":59,"sourceHtmlUrl":60},"metadata.json","REFERENCE","参考文档",458,"https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fnvidia-skills\u002Ftao-train-mask-auto-encoder\u002Fmetadata.json",[62,70,75,82,87,94],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":66,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":19},"01M7D6MC9FAAEV4CMNYX8W7JTG","arxiv-search","Search arXiv physics, math, and computer science preprints using natural language queries. Powered by Valyu semantic search.","MIT","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fyorkeccak-scientific-skills\u002Farxiv-search\u002Fskills\u002Farxiv-search\u002FSKILL.md",[],[],{"skillNo":71,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":66,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"publishedAt":19},"01M7D6M85CKKN87TA98VCQV4H7","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fyorkeccak-scientific-skills\u002Farxiv-search\u002FSkill.md",[],[],{"skillNo":76,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":66,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":19},"01M7D4ZTCS4DR3CE4JY06X84T2","working-with-legacy-code","Safely change and test untested codebases using Feathers' \"Working Effectively with Legacy Code\". Use when the user mentions \"legacy code\", \"no tests\", \"untested codebase\", \"how do I test this\", \"seams\", \"characterization tests\", \"golden master\", \"sprout method\", \"afraid to change this code\", \"monster method\", \"dependency breaking\", or \"inherited a messy codebase\". Also trigger when changing code without tests safely, getting a class under test when constructors, statics, or singletons block it, adding features to tangled modules, or planning incremental test coverage for an old codebase. Covers the legacy-code change algorithm, seams, characterization tests, sprout\u002Fwrap, and dependency-breaking techniques. For refactoring code that already has tests, see refactoring-patterns. For day-to-day code quality, see clean-code.","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fwondelai-skills\u002Fworking-with-legacy-code\u002Fworking-with-legacy-code\u002FSKILL.md",[],[],{"skillNo":83,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":66,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":19},"01M7D4ZHZKJD2RSNGAF63V14P4","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fwondelai-skills\u002Fworking-with-legacy-code\u002FSkill.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":66,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":19},"01M7D4ZB4CJBFW5JE67CPH2QRV","web-typography","Select, pair, and implement typefaces for web projects. Use when the user mentions \"font pairing\", \"which typeface\", \"line height\", \"responsive typography\", \"web font loading\", \"type hierarchy\", \"variable fonts\", \"FOUT\u002FFOIT\", \"typographic scale\", or \"the text is hard to read\". Also trigger when choosing between system fonts and web fonts, optimizing font-loading performance, or designing readable long-form content. Covers readability evaluation, CSS implementation, and performance optimization. For overall UI design systems, see refactoring-ui. For dramatic typographic experiences, see top-design.","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fwondelai-skills\u002Fweb-typography\u002Fweb-typography\u002FSKILL.md",[],[],{"skillNo":95,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":66,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":98,"publishedAt":19},"01M7D4Z4MVYVY3NMJ18BNV6WNA","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fwondelai-skills\u002Fweb-typography\u002FSkill.md",[],[]]