Tao Train Ocdnet

NVIDIAOfficialCreated by NVIDIA Corporation

Skill
Tool integration
Data and AI
Apache-2.0
Medium risk

From the author

OCDNet for scene text detection. Detects arbitrary-oriented text regions in natural images using a differentiable binarization approach. Use when training, evaluating, exporting, pruning, quantizing, retraining, or running inference for a TAO OCDNet model. Trigger phrases include "train OCDNet", "scene text detection", "arbitrary-oriented text boxes", "differentiable binarization detector".

— the description by NVIDIA Corporation

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

The tao-train-ocdnet skill provides a comprehensive workflow for managing the lifecycle of OCDNet models, which are designed for detecting text in natural images using differentiable binarization. This skill supports the full pipeline, including training, pruning, quantization, evaluation, and export. It is integrated with TAO AutoML for optimized training and provides specific guidance for handling datasets, managing checkpoints, and configuring model parameters like backbones and binarization thresholds.

What it does

Trains OCDNet models for scene text detection with support for AutoML.
Supports model pruning and quantization to optimize performance.
Evaluates model accuracy using hmean metrics.
Exports models for deployment and inference.
Manages training resumes and retraining from pruned graphs.
Provides configuration templates for various stages of the model lifecycle.

When to use it

When you need to detect arbitrary-oriented text in natural images.
When building a computer vision pipeline for text extraction.
When you need to optimize a text detection model through pruning or quantization.
When you require an automated approach to hyperparameter tuning for OCDNet training.

Advantages

Supports differentiable binarization for high-precision text detection.
Includes AutoML integration to simplify hyperparameter optimization.
Provides a structured workflow for model lifecycle management from training to deployment.
Offers flexible configuration for multi-GPU and multi-node training environments.

Usage examples

“Train an OCDNet model on my custom dataset.”: The agent initiates the training process using the specified dataset path, applying AutoML policies to optimize the model based on the train_loss metric.
“Prune my trained OCDNet model to reduce its size.”: The agent executes the pruning action using the best checkpoint from the previous training job and the validation dataset to maintain accuracy.
“Evaluate the performance of my OCDNet model.”: The agent runs the evaluation workflow using the provided checkpoint and validation dataset, reporting the hmean metric.
“Quantize my OCDNet model for deployment.”: The agent performs quantization on the model, utilizing calibration images to ensure the model is ready for efficient inference.

Install

Download the skill as a ZIP:
https://autogineer.com/agent-skills/nvidia/tao-train-ocdnet/download

Unzip it into the cross-client skills folder of your project:

mkdir -p .agents/skills && unzip tao-train-ocdnet.zip -d .agents/skills

Documentation by NVIDIA Corporation

OCDNet

Standalone install? If this session was not initialized by the TAO skill bank plugin, run the tao-setup skill first (host preflight, credentials, cross-skill discovery).

OCDNet for scene text detection. Detects arbitrary-oriented text regions in natural images using a differentiable binarization approach.

Set model.pretrained_model_path for pretrained weights.

For TAO Deploy TensorRT actions (gen_trt_engine, TensorRT evaluate, and TensorRT inference), read references/tao-deploy-ocdnet.md first. Deploy spec templates live in this skill's references/ folder with the spec_template_deploy_*.yaml prefix.

The PyT OCDNet CLI supports train, evaluate, export, inference, prune, quantize, and default_specs. It does not expose PyT-side retrain or gen_trt_engine subcommands. The model skill exposes retrain by running ocdnet train with model.load_pruned_graph: true and model.pruned_graph_path. Resume from an epoch checkpoint uses ocdnet train plus train.resume_training_checkpoint_path. TensorRT engine generation is owned by the deploy workflow.

Dataclass Schemas

Generated TAO Core schemas are packaged in schemas/<action>.schema.json, with schemas/manifest.json listing available actions. Each generated schema also emits references/spec_template_<action>.yaml from the schema top-level default field. AutoML enablement is declared at the model layer in references/skill_info.yaml via automl_enabled. Runnable AutoML for an action requires schemas/<action>.schema.json and references/spec_template_<action>.yaml to exist and parse. Use the packaged selected-action schema for automl_default_parameters, automl_disabled_parameters, defaults, min/max bounds, enums, option weights, math conditions, dependencies, and popular parameters. Do not expect ~/tao-core at runtime; maintainers regenerate schemas/templates before packaging the skill bank.

Train Action Policy

This model is AutoML-enabled at the model layer. Before handling any train-stage request, read references/skill_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 / 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/train.schema.json and references/spec_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/application overrides for datasets, specs, output directories, GPU/platform settings, parent checkpoints, and automl_policy. Use direct model training only when automl_policy: off or the packaged train schema/template is missing; in the missing-schema case, report that AutoML is enabled but not runnable for this model until schemas are generated.

For AutoML train, use train_loss_epoch or train_loss as the optimization metric with direction=minimize. The Lightning progress log emits train_loss_epoch, and TAO status.json records the same final value under train_loss. For one-epoch local AutoML smoke runs, set train.lr_scheduler.args.warmup_epoch: 0; leaving warmup equal to the epoch budget causes the trainer to fail before a recommendation can report a metric. Non-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.

Training Requirements

Dataset type: ocdnet
Formats: default
AutoML training metric: train_loss (the status logger's final train_loss_epoch value), with direction=minimize
AutoML metric contract: Use train_loss emitted during training and minimize it. Treat train_loss_epoch as a log fallback alias only; use standalone hmean solely to validate the selected checkpoint.
Standalone evaluation metric: hmean, with direction=maximize

Per-Action Dataset Requirements

Action	Spec Key	Source	Runtime value	List?
evaluate	dataset.validate_dataset.data_path	eval_dataset	extracted validation split folder with `img/` and `gt/`	Yes
inference	inference.input_folder	inference_dataset or eval_dataset	extracted image folder	No
prune	dataset.validate_dataset.data_path	eval_dataset	extracted validation split folder with `img/` and `gt/`	Yes
quantize	dataset.train_dataset.data_path	train_datasets	extracted train split folder with `img/` and `gt/`	Yes
quantize	dataset.validate_dataset.data_path	eval_dataset	extracted validation split folder with `img/` and `gt/`	Yes
quantize	dataset.quant_calibration_dataset.images_dir	train_datasets or calibration_dataset	extracted calibration image folder	No
train	dataset.train_dataset.data_path	train_datasets	extracted train split folder with `img/` and `gt/`	Yes
train	dataset.validate_dataset.data_path	eval_dataset	extracted validation split folder with `img/` and `gt/`	Yes
retrain	dataset.train_dataset.data_path	train_datasets	extracted train split folder with `img/` and `gt/`	Yes
retrain	dataset.validate_dataset.data_path	eval_dataset	extracted validation split folder with `img/` and `gt/`	Yes

Typical Spec Overrides

Data 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. OCDNet does not unpack dataset archives at runtime. If the source is train.tar.gz, test.tar.gz, or img.tar.gz, extract it first and pass the split folder or image folder into the spec. The split folder must contain img/ and gt/; alternatively, pass a UTF-8 datalist text file whose lines map image paths to label paths.

TRAIN_ROOT = "/path/to/extracted/train"
EVAL_ROOT = "/path/to/extracted/test"
INFER_IMG_DIR = "/path/to/extracted/test/img"
CALIB_IMG_DIR = "/path/to/extracted/train/img"

train (mandatory data sources):

{
    "train.num_epochs": 30,
    "train.checkpoint_interval": 10,
    "train.validation_interval": 10,
    "train.num_gpus": 1,
    "dataset.train_dataset.loader.batch_size": 16,
    "dataset.train_dataset.data_path": [TRAIN_ROOT],
    "dataset.validate_dataset.data_path": [EVAL_ROOT],
}

evaluate (mandatory data sources):

{
    "evaluate.checkpoint": "<selected train/AutoML checkpoint>",
    "dataset.validate_dataset.data_path": [EVAL_ROOT],
}

inference (mandatory data sources):

{
    "inference.checkpoint": "<selected train/AutoML checkpoint>",
    "inference.input_folder": INFER_IMG_DIR,
}

prune (mandatory data sources):

{
    "prune.checkpoint": "<selected train/AutoML checkpoint>",
    "dataset.validate_dataset.data_path": [EVAL_ROOT],
}

quantize (mandatory data sources):

{
    "quantize.model_path": "<selected train checkpoint or exported ONNX>",
    "dataset.train_dataset.data_path": [TRAIN_ROOT],
    "dataset.validate_dataset.data_path": [EVAL_ROOT],
    "dataset.quant_calibration_dataset.images_dir": CALIB_IMG_DIR,
}

resume training (mandatory data sources):

{
    "train.resume_training_checkpoint_path": "<exact model_epoch checkpoint>",
    "dataset.train_dataset.data_path": [TRAIN_ROOT],
    "dataset.validate_dataset.data_path": [EVAL_ROOT],
}

retrain from prune output (mandatory data sources):

{
    "model.load_pruned_graph": True,
    "model.pruned_graph_path": "<selected prune output>",
    "dataset.train_dataset.data_path": [TRAIN_ROOT],
    "dataset.validate_dataset.data_path": [EVAL_ROOT],
}

default_specs:

{
    "results_dir": "<writable output directory>",
}

Eval Dataset

Optional. Test dataset provided as separate tarball.

Important Parameters

model.backbone: Default deformable_resnet18. Deformable convolutions improve text region detection for irregular text.
train.optimizer.args.lr: Learning rate. Default 0.001 (Adam).
postprocess.thresh: Binarization threshold for text region extraction.
postprocess.box_thresh: Box confidence threshold for filtering detections.

Multi-GPU / Multi-Node

Launch method: Lightning-managed (single python process, Lightning spawns workers).

Spec Key	Description	Default
`train.num_gpus`	Number of GPUs	1
`train.gpu_ids`	GPU device indices	[0]
`train.distributed_strategy`	`ddp`, `fsdp`, or `deepspeed_stage_3_offload`	`ddp`

ddp with activation checkpointing: find_unused_parameters=False
ddp without: find_unused_parameters=True
fsdp forces FP16
deepspeed_stage_3_offload is uniquely supported for OCDNet (forces FP16)
FAN backbones auto-enable sync_batchnorm

Hardware

Minimum 1 GPU(s), recommended 1 GPU(s). 8GB+ VRAM per GPU. OCDNet is lightweight. Single GPU is sufficient for most datasets.

Error Patterns

Low detection rate: Tune postprocess.thresh and box_thresh. Default thresholds may be too aggressive for some datasets.

One-epoch smoke train with default scheduler: train.num_epochs must not equal train.lr_scheduler.args.warmup_epoch. For one-epoch validation, set warmup_epoch: 0; for normal starter runs, keep num_epochs > warmup_epoch.

Archive passed as dataset path: dataset.*.data_path is not an archive path for OCDNet. Passing train.tar.gz or test.tar.gz directly causes the dataloader to open the gzip as a UTF-8 datalist. Extract the archive and pass the split folder containing img/ and gt/, or pass a real UTF-8 datalist file.

Quantize checkpoint type: Do not pass model_best.pth to the PyTorch quantize path. Some older PyT runtimes wrote model_best.pth without full Lightning checkpoint metadata. The default torchao quantize path should use the intended full model_epoch_<epoch>_step_<step>.pth checkpoint and write quantized_model_torchao.pth.

Default specs output directory: ocdnet default_specs requires a writable results_dir override, for example results_dir=/workspace/run/results/default_specs.

Checkpoint Handoff

OCDNet train writes model_best.pth plus full Lightning epoch checkpoints such as model_epoch_001_step_00046.pth; it may also write ocd_model_latest.pth as a latest symlink. Use model_best.pth for evaluate.checkpoint, inference.checkpoint, export.checkpoint, and prune.checkpoint when the user asks for the best checkpoint. Use a specific model_epoch_<epoch>_step_<step>.pth for train.resume_training_checkpoint_path and for any action that explicitly needs a full Lightning checkpoint. Prune writes artifacts such as pruned_<ch_sparsity>.pth; use the exact pruned .pth artifact for model.pruned_graph_path when retraining from a pruned graph. Use a latest checkpoint only when the user explicitly asks for latest.

If quantize is retried with a PyTorch backend, resolve the full model_epoch_<epoch>_step_<step>.pth that corresponds to the intended best epoch or requested epoch; do not pass model_best.pth to the PyTorch quantize path. If quantize is retried with modelopt.onnx, pass the exported ONNX as quantize.model_path and verify that the runtime image actually contains modelopt.onnx.quantization.

Spec Param / Parent Model Inference

Model-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.

Model handoff mappings:

Action	Spec Field	Inference Function	Meaning
evaluate	`evaluate.checkpoint`	`parent_model`	model file inferred from the parent job results folder
evaluate	`results_dir`	`output_dir`	current job results directory
export	`export.checkpoint`	`parent_model`	model file inferred from the parent job results folder
export	`export.onnx_file`	`create_onnx_file`	output ONNX path
export	`results_dir`	`output_dir`	current job results directory
inference	`inference.checkpoint`	`parent_model`	model file inferred from the parent job results folder
inference	`results_dir`	`output_dir`	current job results directory
prune	`prune.checkpoint`	`parent_model`	model file inferred from the parent job results folder
prune	`results_dir`	`output_dir`	current job results directory
quantize	`quantize.model_path`	`parent_model`	model file inferred from the parent job results folder
quantize	`results_dir`	`output_dir`	current job results directory
retrain from prune	`model.pruned_graph_path`	`parent_model`	exact pruned model file inferred from the parent prune results folder
retrain from prune	`results_dir`	`output_dir`	current job results directory
train	`model.pretrained_model_path`	`ptm_if_no_resume_model`	PTM when no resume checkpoint exists
train	`results_dir`	`output_dir`	current job results directory
train	`train.resume_training_checkpoint_path`	`resume_model`	model file inferred from the current job results folder

For parent_model or parent_model_folder, pass the upstream train/export/AutoML 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.

Deployment

tao-deploy-ocdnet

Files

ATTRIBUTION.mdAdded by AutoGineer918 B
BENCHMARK.md6.9 KiB
LICENSELicence file added by AutoGineer11.2 KiB
SKILL.md14.2 KiB
config/skillspector-baseline.yaml834 B
evals/evals.json760 B
references/skill_info.yaml6.1 KiB
references/spec_template_deploy_evaluate.yaml814 B
references/spec_template_deploy_gen_trt_engine.yaml445 B
references/spec_template_deploy_inference.yaml413 B
references/spec_template_evaluate.yaml3.3 KiB
references/spec_template_export.yaml3.1 KiB
references/spec_template_gen_trt_engine.yaml3.4 KiB
references/spec_template_inference.yaml3.3 KiB
references/spec_template_prune.yaml3.0 KiB
references/spec_template_quantize.yaml2.9 KiB
references/spec_template_retrain.yaml2.9 KiB
references/spec_template_train.yaml2.9 KiB
references/tao-deploy-ocdnet.md6.2 KiB
references/tao-deploy-ocdnet.skill_info.yaml3.3 KiB
schemas/evaluate.schema.json60.7 KiB
schemas/export.schema.json55.7 KiB
schemas/gen_trt_engine.schema.json62.6 KiB
schemas/inference.schema.json60.7 KiB
schemas/manifest.json26.2 KiB
schemas/prune.schema.json55.3 KiB
schemas/quantize.schema.json53.3 KiB
schemas/retrain.schema.json53.3 KiB
schemas/train.schema.json53.3 KiB
skill-card.md4.4 KiB
skill.oms.sig10.3 KiB

schemas/quantize.schema.json

{
  "automl_default_parameters": [
    "train.optimizer.args.weight_decay",
    "train.optimizer.args.lr"
  ],
  "automl_disabled_parameters": [
    "quantize",
    "train.cudnn",
    "train.post_processing.args",
    "gen_trt_engine.tensorrt.calibration.cal_image_dir",
    "quantize.backend_kwargs",
    "dataset.train_dataset.args",
    "train.gpu_ids",
    "dataset.validate_dataset",
    "dataset.validate_dataset.loader",
    "wandb.tags",
    "dataset.validate_dataset.data_path",
    "train.metric",
    "train.optimizer.args",
    "dataset.train_dataset.data_path",
    "quantize.skip_names",
    "evaluate.post_processing",
    "dataset.train_dataset",
    "evaluate",
    "evaluate.metric",
    "evaluate.metric.args",
    "inference",
    "train",
    "evaluate.post_processing.args",
    "gen_trt_engine",
    "train.post_processing",
    "gen_trt_engine.tensorrt.layers_precision",
    "dataset.train_dataset.args.pre_processes",
    "dataset",
    "dataset.validate_dataset.args.ignore_tags",
    "gen_trt_engine.tensorrt",
    "quantize.layers",
    "dataset.train_dataset.loader",
    "dataset.quant_calibration_dataset",
    "dataset.train_dataset.loader.batch_size",
    "dataset.validate_dataset.args.filter_keys",
    "dataset.train_dataset.args.filter_keys",
    "train.lr_scheduler.args",
    "train.loss",
    "model",
    "inference.post_processing",
    "evaluate.gpu_ids",
    "gen_trt_engine.tensorrt.calibration",
    "train.trainer",
    "dataset.validate_dataset.args.pre_processes",
    "train.lr_scheduler",
    "train.metric.args",
    "export",
    "wandb",
    "inference.post_processing.args",
    "inference.gpu_ids",
    "prune",
    "dataset.validate_dataset.args",
    "train.optimizer",
    "dataset.train_dataset.args.ignore_tags"
  ],
  "default": {
    "dataset": {
      "quant_calibration_dataset": {
        "images_dir": ""
      },
      "train_dataset": {
        "args": {
          "filter_keys": [
            "img_path",
            "img_name",
            "text_polys",
            "texts",
            "ignore_tags",
            "shape"
          ],
          "ignore_tags": [
            "*",
            "###"
          ],
          "img_mode": "BGR",
          "pre_processes": [
            {
              "args": {
                "keep_ratio": true,
                "max_tries": 50,
                "size": [
                  640,
                  640
                ]
              },
              "type": "EastRandomCropData"
            },
            {
              "args": {
                "shrink_ratio": 0.4,
                "thresh_max": 0.7,
                "thresh_min": 0.3
              },
              "type": "MakeBorderMap"
            },
            {
              "args": {
                "min_text_size": 8,
                "shrink_ratio": 0.4
              },
              "type": "MakeShrinkMap"
            }
          ]
        },
        "data_name": "ICDAR2015Dataset",
        "data_path": [],
        "loader": {
          "batch_size": 16,
          "collate_fn": "",
          "num_workers": 0,
          "pin_memory": false,
          "shuffle": true
        }
      },
      "validate_dataset": {
        "args": {
          "filter_keys": [
            ""
          ],
          "ignore_tags": [
            "*",
            "###"
          ],
          "img_mode": "BGR",
          "pre_processes": [
            {
              "args": {
                "resize_text_polys": true,
                "short_size": [
                  1280,
                  736
                ]
              },
              "type": "Resize2D"
            }
          ]
        },
        "data_name": "ICDAR2015Dataset",
        "data_path": [],
        "loader": {
          "batch_size": 1,
          "collate_fn": "ICDARCollateFN",
          "num_workers": 0,
          "pin_memory": false,
          "shuffle": false
        }
      }
    },
    "encryption_key": "",
    "model": {
      "activation_checkpoint": false,
      "backbone": "deformable_resnet18",
      "enlarge_feature_map_size": false,
      "fuse_qkv_proj": true,
      "head": "DBHead",
      "in_channels": 3,
      "inner_channels": 256,
      "k": 50,
      "load_pruned_graph": false,
      "neck": "FPN",
      "out_channels": 2,
      "pretrained": false,
      "pretrained_model_path": "",
      "pruned_graph_path": "",
      "quant": false
    },
    "model_name": "",
    "quantize": {
      "algorithm": "minmax",
      "backend": "torchao",
      "backend_kwargs": {},
      "device": "cuda",
      "layers": [],
      "mode": "weight_only_ptq",
      "model_path": "",
      "results_dir": "",
      "skip_names": []
    },
    "results_dir": "",
    "train": {
      "checkpoint_interval": 1,
      "checkpoint_interval_unit": "epoch",
      "cudnn": {
        "benchmark": false,
        "deterministic": true
      },
      "distributed_strategy": "ddp",
      "gpu_ids": [
        0
      ],
      "is_dry_run": false,
      "loss": {
        "alpha": 5,
        "beta": 10,
        "eps": 1e-06,
        "ohem_ratio": 3,
        "type": "DBLoss"
      },
      "lr_scheduler": {
        "args": {
          "warmup_epoch": 3
        },
        "type": "WarmupPolyLR"
      },
      "metric": {
        "args": {
          "is_output_polygon": false
        },
        "type": "QuadMetric"
      },
      "model_ema": false,
      "model_ema_decay": 0.9999,
      "num_epochs": 10,
      "num_gpus": 1,
      "num_nodes": 1,
      "optimizer": {
        "args": {
          "amsgrad": true,
          "eps": 1e-08,
          "lr": 0.001,
          "momentum": 0.0,
          "weight_decay": 0.0
        },
        "type": "Adam"
      },
      "post_processing": {
        "args": {
          "box_thresh": 0.55,
          "max_candidates": 1000,
          "thresh": 0.3,
          "unclip_ratio": 1.5
        },
        "type": "SegDetectorRepresenter"
      },
      "precision": "fp32",
      "results_dir": "",
      "resume_training_checkpoint_path": "",
      "seed": 1234,
      "trainer": {
        "clip_grad_norm": 5.0
      },
      "use_distributed_sampler": false,
      "validation_interval": 1
    },
    "wandb": {
      "enable": true,
      "entity": "",
      "group": "",
      "name": "TAO Toolkit Training",
      "project": "TAO Toolkit",
      "reinit": false,
      "run_id": "",
      "save_code": false,
      "sync_tensorboard": false,
      "tags": [
        "tao-toolkit"
      ]
    }
  },
  "popular": {
    "dataset": {
      "train_dataset": {
        "loader": {
          "batch_size": 16,
          "num_workers": 0
        }
      },
      "validate_dataset": {
        "loader": {
          "batch_size": 1,
          "num_workers": 0
        }
      }
    },
    "evaluate": {
      "batch_size": 1,
      "gpu_ids": [
        0
      ],
      "num_gpus": 1,
      "num_nodes": 1,
      "post_processing": {
        "args": {
          "box_thresh": 0.55,
          "max_candidates": 1000,
          "thresh": 0.3,
          "unclip_ratio": 1.5
        }
      }
    },
    "export": {
      "height": 736,
      "opset_version": 11,
      "width": 1280
    },
    "gen_trt_engine": {
      "batch_size": -1,
      "gpu_id": 0,
      "height": 736,
      "tensorrt": {
        "calibration": {
          "cal_batch_size": 1,
          "cal_batches": 1
        },
        "max_batch_size": 1,
        "min_batch_size": 1,
        "opt_batch_size": 1
      },
      "width": 1280
    },
    "inference": {
      "gpu_ids": [
        0
      ],
      "height": 736,
      "num_gpus": 1,
      "num_nodes": 1,
      "post_processing": {
        "args": {
          "box_thresh": 0.55,
          "max_candidates": 1000,
          "thresh": 0.3,
          "unclip_ratio": 1.5
        }
      },
      "width": 1280
    },
    "model": {
      "backbone": "deformable_resnet18"
    },
    "train": {
      "checkpoint_interval": 1,
      "gpu_ids": [
        0
      ],
      "num_epochs": 10,
      "num_gpus": 1,
      "num_nodes": 1,
      "optimizer": {
        "args": {
          "lr": 0.001
        }
      },
      "post_processing": {
        "args": {
          "box_thresh": 0.55,
          "max_candidates": 1000,
          "thresh": 0.3,
          "unclip_ratio": 1.5
        }
      },
      "validation_interval": 1
    }
  },
  "properties": {
    "automl_disabled_parameters": [
      "wandb",
      "model",
      "dataset",
      "train",
      "evaluate",
      "inference",
      "export",
      "gen_trt_engine",
      "prune",
      "quantize"
    ],
    "dataset": {
      "automl_disabled_parameters": [
        "dataset.train_dataset",
        "dataset.validate_dataset",
        "dataset.quant_calibration_dataset"
      ],
      "automl_enabled": false,
      "default": {
        "quant_calibration_dataset": {
          "images_dir": ""
        },
        "train_dataset": {
          "args": {
            "filter_keys": [
              "img_path",
              "img_name",
              "text_polys",
              "texts",
              "ignore_tags",
              "shape"
            ],
            "ignore_tags": [
              "*",
              "###"
            ],
            "img_mode": "BGR",
            "pre_processes": [
              {
                "args": {
                  "keep_ratio": true,
                  "max_tries": 50,
                  "size": [
                    640,
                    640
                  ]
                },
                "type": "EastRandomCropData"
              },
              {
                "args": {
                  "shrink_ratio": 0.4,
                  "thresh_max": 0.7,
                  "thresh_min": 0.3
                },
                "type": "MakeBorderMap"
              },
              {
                "args": {
                  "min_text_size": 8,
                  "shrink_ratio": 0.4
                },
                "type": "MakeShrinkMap"
              }
            ]
          },
          "data_name": "ICDAR2015Dataset",
          "data_path": [],
          "loader": {
            "batch_size": 16,
            "collate_fn": "",
            "num_workers": 0,
            "pin_memory": false,
            "shuffle": true
          }
        },
        "validate_dataset": {
          "args": {
            "filter_keys": [
              ""
            ],
            "ignore_tags": [
              "*",
              "###"
            ],
            "img_mode": "BGR",
            "pre_processes": [
              {
                "args": {
                  "resize_text_polys": true,
                  "short_size": [
                    1280,
                    736
                  ]
                },
                "type": "Resize2D"
              }
            ]
          },
          "data_name": "ICDAR2015Dataset",
          "data_path": [],
          "loader": {
            "batch_size": 1,
            "collate_fn": "ICDARCollateFN",
            "num_workers": 0,
            "pin_memory": false,
            "shuffle": false
          }
        }
      },
      "description": "Configurable parameters to construct the dataset for an OCDNet experiment.",
      "popular": [
        "validate_dataset",
        "train_dataset"
      ],
      "properties": {
        "quant_calibration_dataset": {
          "automl_enabled": false,
          "default": {
            "images_dir": ""
          },
          "description": "Configurable parameters for quantization calibration dataset.",
          "properties": {
            "images_dir": {
              "default": "",
              "description": "Path to the directory containing calibration images.",
              "title": "calibration images directory",
              "type": "string"
            }
          },
          "type": "collection"
        },
        "train_dataset": {
          "automl_disabled_parameters": [
            "dataset.train_dataset.data_path",
            "dataset.train_dataset.args",
            "dataset.train_dataset.loader"
          ],
          "automl_enabled": false,
          "default": {
            "args": {
              "filter_keys": [
                "img_path",
                "img_name",
                "text_polys",
                "texts",
                "ignore_tags",
                "shape"
              ],
              "ignore_tags": [
                "*",
                "###"
              ],
              "img_mode": "BGR",
              "pre_processes": [
                {
                  "args": {
                    "keep_ratio": true,
                    "max_tries": 50,
                    "size": [
                      640,
                      640
                    ]
                  },
                  "type": "EastRandomCropData"
                },
                {
                  "args": {
                    "shrink_ratio": 0.4,
                    "thresh_max": 0.7,
                    "thresh_min": 0.3
                  },
                  "type": "MakeBorderMap"
                },
                {
                  "args": {
                    "min_text_size": 8,
                    "shrink_ratio": 0.4
                  },
                  "type": "MakeShrinkMap"
                }
              ]
            },
            "data_name": "ICDAR2015Dataset",
            "data_path": [],
            "loader": {
              "batch_size": 16,
              "collate_fn": "",
              "num_workers": 0,
              "pin_memory": false,
              "shuffle": true
            }
          },
          "description": "Hyper parameters to configure the training dataset.",
          "popular": [
            "loader"
          ],
          "properties": {
            "args": {
              "automl_disabled_parameters": [
                "dataset.train_dataset.args.filter_keys",
                "dataset.train_dataset.args.ignore_tags",
                "dataset.train_dataset.args.pre_processes"
              ],
              "automl_enabled": false,
              "default": {
                "filter_keys": [
                  "img_path",
                  "img_name",
                  "text_polys",
                  "texts",
                  "ignore_tags",
                  "shape"
                ],
                "ignore_tags": [
                  "*",
                  "###"
                ],
                "img_mode": "BGR",
                "pre_processes": [
                  {
                    "args": {
                      "keep_ratio": true,
                      "max_tries": 50,
                      "size": [
                        640,
                        640
                      ]
                    },
                    "type": "EastRandomCropData"
                  },
                  {
                    "args": {
                      "shrink_ratio": 0.4,
                      "thresh_max": 0.7,
                      "thresh_min": 0.3
                    },
                    "type": "MakeBorderMap"
                  },
                  {
                    "args": {
                      "min_text_size": 8,
                      "shrink_ratio": 0.4
                    },
                    "type": "MakeShrinkMap"
                  }
                ]
              },
              "description": "Configurable parameters to construct the training dataset.",
              "properties": {
                "filter_keys": {
                  "automl_enabled": false,
                  "default": [
                    "img_path",
                    "img_name",
                    "text_polys",
                    "texts",
                    "ignore_tags",
                    "shape"
                  ],
                  "description": "List of ignored keys",
                  "title": "filter_keys",
                  "type": "list"
                },
                "ignore_tags": {
                  "automl_enabled": false,
                  "default": [
                    "*",
                    "###"
                  ],
                  "description": "List of labels that are not used to train",
                  "title": "ignore_tags",
                  "type": "list"
                },
                "img_mode": {
                  "default": "BGR",
                  "description": "The image mode.",
                  "enum": [
                    "BGR",
                    "RGB",
                    "GRAY"
                  ],
                  "title": "img_mode",
                  "type": "categorical"
                },
                "pre_processes": {
                  "automl_enabled": false,
                  "default": [
                    {
                      "args": {
                        "keep_ratio": true,
                        "max_tries": 50,
                        "size": [
                          640,
                          640
                        ]
                      },
                      "type": "EastRandomCropData"
                    },
                    {
                      "args": {
                        "shrink_ratio": 0.4,
                        "thresh_max": 0.7,
                        "thresh_min": 0.3
                      },
                      "type": "MakeBorderMap"
                    },
                    {
                      "args": {
                        "min_text_size": 8,
                        "shrink_ratio": 0.4
                      },
                      "type": "MakeShrinkMap"
                    }
                  ],
                  "description": "The pre-processing configuration.",
                  "title": "pre_processes",
                  "type": "list"
                }
              },
              "type": "collection"
            },
            "data_name": {
              "default": "ICDAR2015Dataset",
              "description": "The dataset type",
              "title": "data_name",
              "type": "string"
            },
            "data_path": {
              "automl_enabled": false,
              "default": [],
              "description": "The list of training dataset paths",
              "title": "data_path",
              "type": "list"
            },
            "loader": {
              "automl_disabled_parameters": [
                "dataset.train_dataset.loader.batch_size"
              ],
              "automl_enabled": false,
              "default": {
                "batch_size": 16,
                "collate_fn": "",
                "num_workers": 0,
                "pin_memory": false,
                "shuffle": true
              },
              "description": "Configurable parameters to construct the training dataloader.",
              "popular": [
                "batch_size",
                "num_workers"
              ],
              "properties": {
                "batch_size": {
                  "automl_enabled": false,
                  "default": 16,
                  "description": "The batch size during training.",
                  "maximum": Infinity,
                  "minimum": 1,
                  "popular": true,
                  "title": "batch_size",
                  "type": "int"
                },
                "collate_fn": {
                  "default": "",
                  "description": "The collate function.",
                  "type": "string"
                },
                "num_workers": {
                  "default": 0,
                  "description": "The threads used to load data.",
                  "maximum": Infinity,
                  "minimum": 0,
                  "popular": true,
                  "title": "num_workers",
                  "type": "int"
                },
                "pin_memory": {
                  "default": false,
                  "description": "Flag to enable pinned memory or not",
                  "title": "pin_memory",
                  "type": "bool"
                },
                "shuffle": {
                  "default": true,
                  "description": "Flag to shuffle the data or not.",
                  "title": "shuffle",
                  "type": "bool"
                }
              },
              "type": "collection"
            }
          },
          "title": "train_dataset",
          "type": "collection"
        },
        "validate_dataset": {
          "automl_disabled_parameters": [
            "dataset.validate_dataset.data_path",
            "dataset.validate_dataset.args",
            "dataset.validate_dataset.loader"
          ],
          "automl_enabled": false,
          "default": {
            "args": {
              "filter_keys": [
                ""
              ],
              "ignore_tags": [
                "*",
                "###"
              ],
              "img_mode": "BGR",
              "pre_processes": [
                {
                  "args": {
                    "resize_text_polys": true,
                    "short_size": [
                      1280,
                      736
                    ]
                  },
                  "type": "Resize2D"
                }
              ]
            },
            "data_name": "ICDAR2015Dataset",
            "data_path": [],
            "loader": {
              "batch_size": 1,
              "collate_fn": "ICDARCollateFN",
              "num_workers": 0,
              "pin_memory": false,
              "shuffle": false
            }
          },
          "description": "Hyper parameters to configure the validation dataset.",
          "popular": [
            "loader"
          ],
          "properties": {
            "args": {
              "automl_disabled_parameters": [
                "dataset.validate_dataset.args.filter_keys",
                "dataset.validate_dataset.args.ignore_tags",
                "dataset.validate_dataset.args.pre_processes"
              ],
              "automl_enabled": false,
              "default": {
                "filter_keys": [
                  ""
                ],
                "ignore_tags": [
                  "*",
                  "###"
                ],
                "img_mode": "BGR",
                "pre_processes": [
                  {
                    "args": {
                      "resize_text_polys": true,
                      "short_size": [
                        1280,
                        736
                      ]
                    },
                    "type": "Resize2D"
                  }
                ]
              },
              "description": "Configurable parameters to construct the validation dataset.",
              "properties": {
                "filter_keys": {
                  "automl_enabled": false,
                  "default": [
                    ""
                  ],
                  "description": "List of ignored keys",
                  "title": "filter_keys",
                  "type": "list"
                },
                "ignore_tags": {
                  "automl_enabled": false,
                  "default": [
                    "*",
                    "###"
                  ],
                  "description": "List of labels that are not used to evaluate",
                  "title": "ignore_tags",
                  "type": "list"
                },
                "img_mode": {
                  "default": "BGR",
                  "description": "The image mode.",
                  "enum": [
                    "BGR",
                    "RGB",
                    "GRAY"
                  ],
                  "title": "img_mode",
                  "type": "categorical"
                },
                "pre_processes": {
                  "automl_enabled": false,
                  "default": [
                    {
                      "args": {
                        "resize_text_polys": true,
                        "short_size": [
                          1280,
                          736
                        ]
                      },
                      "type": "Resize2D"
                    }
                  ],
                  "description": "The pre-processing configuration.",
                  "title": "pre_processes",
                  "type": "list"
                }
              },
              "type": "collection"
            },
            "data_name": {
              "default": "ICDAR2015Dataset",
              "description": "The dataset type",
              "title": "data_name",
              "type": "string"
            },
            "data_path": {
              "automl_enabled": false,
              "default": [],
              "description": "The list of training dataset paths",
              "title": "data_path",
              "type": "list"
            },
            "loader": {
              "automl_enabled": false,
              "default": {
                "batch_size": 1,
                "collate_fn": "ICDARCollateFN",
                "num_workers": 0,
                "pin_memory": false,
                "shuffle": false
              },
              "description": "Configurable parameters to construct the validation dataloader.",
              "popular": [
                "batch_size",
                "num_workers"
              ],
              "properties": {
                "batch_size": {
                  "default": 1,
                  "description": "The batch size during evaluation",
                  "maximum": Infinity,
                  "minimum": 1,
                  "popular": true,
                  "title": "batch_size",
                  "type": "int"
                },
                "collate_fn": {
                  "default": "ICDARCollateFN",
                  "description": "The collate function.",
                  "type": "string"
                },
                "num_workers": {
                  "default": 0,
                  "description": "The threads used to load data.",
                  "maximum": Infinity,
                  "minimum": 0,
                  "popular": true,
                  "title": "num_workers",
                  "type": "int"
                },
                "pin_memory": {
                  "default": false,
                  "description": "Flag to enable pinned memory or not",
                  "title": "pin_memory",
                  "type": "bool"
                },
                "shuffle": {
                  "default": false,
                  "description": "Flag to shuffle the data or not",
                  "title": "shuffle",
                  "type": "bool"
                }
              },
              "type": "collection"
            }
          },
          "title": "validate_dataset",
          "type": "collection"
        }
      },
      "type": "collection"
    },
    "encryption_key": {
      "default": "",
      "description": "Key for encrypting model checkpoints",
      "title": "Encryption key",
      "type": "string"
    },
    "model": {
      "automl_enabled": false,
      "default": {
        "activation_checkpoint": false,
        "backbone": "deformable_resnet18",
        "enlarge_feature_map_size": false,
        "fuse_qkv_proj": true,
        "head": "DBHead",
        "in_channels": 3,
        "inner_channels": 256,
        "k": 50,
        "load_pruned_graph": false,
        "neck": "FPN",
        "out_channels": 2,
        "pretrained": false,
        "pretrained_model_path": "",
        "pruned_graph_path": "",
        "quant": false
      },
      "description": "Configurable parameters to construct the model for an OCDNet experiment.",
      "popular": [
        "backbone"
      ],
      "properties": {
        "activation_checkpoint": {
          "default": false,
          "description": "Flag to use activation checkpoints to save GPU memory, only for the FAN-tiny backbone",
          "title": "activation_checkpoint",
          "type": "bool"
        },
        "backbone": {
          "default": "deformable_resnet18",
          "description": "The backbone name of the model.\n                    It supports deformable_resnet18, deformable_resnet50 and fan_tiny_8_p4_hybrid.",
          "enum": [
            "deformable_resnet18",
            "deformable_resnet50",
            "fan_tiny_8_p4_hybrid"
          ],
          "popular": true,
          "title": "backbone",
          "type": "categorical"
        },
        "enlarge_feature_map_size": {
          "default": false,
          "description": "Flag to enlarge the output feature map size of the FAN-tiny backbone",
          "title": "enlarge_feature_map_size",
          "type": "bool"
        },
        "fuse_qkv_proj": {
          "default": true,
          "description": "Flag to fuse the qkv projection",
          "title": "fuse_qkv_proj",
          "type": "bool"
        },
        "head": {
          "default": "DBHead",
          "description": "Head name of the model.",
          "title": "head",
          "type": "string"
        },
        "in_channels": {
          "default": 3,
          "description": "Number of input channels in FPN",
          "maximum": 3,
          "minimum": 3,
          "title": "in_channels",
          "type": "int"
        },
        "inner_channels": {
          "default": 256,
          "description": "Number of inner channels in FPN",
          "maximum": 256,
          "minimum": 256,
          "title": "inner_channels",
          "type": "int"
        },
        "k": {
          "default": 50,
          "description": "Coefficient of Differentiable Binarization",
          "maximum": Infinity,
          "minimum": 1,
          "title": "k",
          "type": "int"
        },
        "load_pruned_graph": {
          "default": false,
          "description": "Flag to load pruned model or not.",
          "title": "load_pruned_graph",
          "type": "bool"
        },
        "neck": {
          "default": "FPN",
          "description": "Neck name of the model.",
          "enum": [
            "FPN",
            "FANNeck"
          ],
          "title": "neck",
          "type": "categorical"
        },
        "out_channels": {
          "default": 2,
          "description": "Number of out channels",
          "maximum": 2,
          "minimum": 2,
          "title": "out_channels",
          "type": "int"
        },
        "pretrained": {
          "default": false,
          "description": "Flag to use pretrained model or not.",
          "title": "pretrained",
          "type": "bool"
        },
        "pretrained_model_path": {
          "default": "",
          "description": "[Optional] Path to a pretrained model file.",
          "title": "pretrained model path",
          "type": "string"
        },
        "pruned_graph_path": {
          "default": "",
          "description": "[Optional] Path to a pruned model file.",
          "title": "pruned model path",
          "type": "string"
        },
        "quant": {
          "default": false,
          "description": "Flag to do quantization",
          "title": "quant",
          "type": "bool"
        }
      },
      "type": "collection"
    },
    "model_name": {
      "default": "",
      "description": "Name of model if invoking task via :code:`model_agnostic`",
      "title": "Model name",
      "type": "string"
    },
    "quantize": {
      "automl_disabled_parameters": [
        "quantize.layers",
        "quantize.skip_names",
        "quantize.backend_kwargs"
      ],
      "automl_enabled": false,
      "default": {
        "algorithm": "minmax",
        "backend": "torchao",
        "backend_kwargs": {},
        "device": "cuda",
        "layers": [],
        "mode": "weight_only_ptq",
        "model_path": "",
        "results_dir": "",
        "skip_names": []
      },
      "description": "Configurable parameters for model quantization.",
      "properties": {
        "algorithm": {
          "default": "minmax",
          "description": "Calibration/optimization algorithm. Used by ModelOpt backends (modelopt.pytorch and modelopt.onnx). Ignored by torchao backend.",
          "enum": [
            "minmax",
            "max",
            "entropy",
            "awq_clip",
            "awq_lite",
            "awq_full",
            "rtn_dq"
          ],
          "title": "Calibration algorithm",
          "type": "categorical"
        },
        "backend": {
          "default": "torchao",
          "description": "The quantization backend to use",
          "enum": [
            "modelopt.pytorch",
            "torchao",
            "modelopt.onnx"
          ],
          "title": "Quantization backend",
          "type": "categorical"
        },
        "backend_kwargs": {
          "automl_enabled": false,
          "description": "Additional keyword arguments to pass to the backend",
          "title": "Backend kwargs",
          "type": "collection"
        },
        "device": {
          "default": "cuda",
          "description": "Device to use for calibration. Accepts 'cuda' (uses default GPU), 'cpu', 'trt' (TensorRT for ONNX backend), or specific GPU device like 'cuda:0', 'cuda:1', etc. If 'cuda' is specified but no GPU is available, the framework will automatically fall back to 'cpu'. Note: 'trt' is only supported by the modelopt.onnx backend for ONNX Runtime with TensorRT execution provider.",
          "pattern": "^(cuda|cpu|trt|cuda:[0-9]+)$",
          "title": "Calibration device",
          "type": "string"
        },
        "layers": {
          "automl_enabled": false,
          "default": [],
          "description": "List of per-module quantization configurations. Each entry specifies which modules to quantize and their data types. This is the primary way to configure quantization.",
          "title": "Layer quantization configs",
          "type": "list"
        },
        "mode": {
          "default": "weight_only_ptq",
          "description": "The quantization mode to use",
          "enum": [
            "static_ptq",
            "weight_only_ptq"
          ],
          "title": "Quantization mode",
          "type": "categorical"
        },
        "model_path": {
          "default": "",
          "description": "Path to the model to be quantized. For ONNX backend, path to ONNX file.",
          "title": "Model path",
          "type": "string"
        },
        "results_dir": {
          "default": "",
          "description": "Path to where all the assets generated from a task are stored.",
          "title": "Results directory",
          "type": "string"
        },
        "skip_names": {
          "automl_enabled": false,
          "default": [],
          "description": "List of module/layer names or patterns to exclude from quantization",
          "title": "Skip names",
          "type": "list"
        }
      },
      "required": [
        "model_path",
        "results_dir"
      ],
      "type": "collection"
    },
    "results_dir": {
      "default": "",
      "description": "\n        Path to where all the assets generated from a task are stored.\n        ",
      "title": "Results directory",
      "type": "string"
    },
    "train": {
      "automl_disabled_parameters": [
        "train.gpu_ids",
        "train.cudnn",
        "train.post_processing",
        "train.metric",
        "train.trainer",
        "train.loss",
        "train.optimizer",
        "train.lr_scheduler"
      ],
      "automl_enabled": false,
      "default": {
        "checkpoint_interval": 1,
        "checkpoint_interval_unit": "epoch",
        "cudnn": {
          "benchmark": false,
          "deterministic": true
        },
        "distributed_strategy": "ddp",
        "gpu_ids": [
          0
        ],
        "is_dry_run": false,
        "loss": {
          "alpha": 5,
          "beta": 10,
          "eps": 1e-06,
          "ohem_ratio": 3,
          "type": "DBLoss"
        },
        "lr_scheduler": {
          "args": {
            "warmup_epoch": 3
          },
          "type": "WarmupPolyLR"
        },
        "metric": {
          "args": {
            "is_output_polygon": false
          },
          "type": "QuadMetric"
        },
        "model_ema": false,
        "model_ema_decay": 0.9999,
        "num_epochs": 10,
        "num_gpus": 1,
        "num_nodes": 1,
        "optimizer": {
          "args": {
            "amsgrad": true,
            "eps": 1e-08,
            "lr": 0.001,
            "momentum": 0.0,
            "weight_decay": 0.0
          },
          "type": "Adam"
        },
        "post_processing": {
          "args": {
            "box_thresh": 0.55,
            "max_candidates": 1000,
            "thresh": 0.3,
            "unclip_ratio": 1.5
          },
          "type": "SegDetectorRepresenter"
        },
        "precision": "fp32",
        "results_dir": "",
        "resume_training_checkpoint_path": "",
        "seed": 1234,
        "trainer": {
          "clip_grad_norm": 5.0
        },
        "use_distributed_sampler": false,
        "validation_interval": 1
      },
      "description": "Configurable parameters to construct the trainer for an OCDNet experiment.",
      "popular": [
        "post_processing",
        "validation_interval",
        "checkpoint_interval",
        "num_gpus",
        "num_nodes",
        "num_epochs",
        "optimizer",
        "gpu_ids"
      ],
      "properties": {
        "checkpoint_interval": {
          "default": 1,
          "description": "The interval (in epochs) at which a checkpoint will be saved. Helps resume training.",
          "minimum": 1,
          "popular": true,
          "title": "Checkpoint interval",
          "type": "int"
        },
        "checkpoint_interval_unit": {
          "default": "epoch",
          "description": "The unit of the checkpoint interval.",
          "enum": [
            "epoch",
            "step"
          ],
          "title": "Checkpoint interval unit",
          "type": "categorical"
        },
        "cudnn": {
          "automl_enabled": false,
          "default": {
            "benchmark": false,
            "deterministic": true
          },
          "properties": {
            "benchmark": {
              "default": false,
              "type": "bool"
            },
            "deterministic": {
              "default": true,
              "type": "bool"
            }
          },
          "type": "collection"
        },
        "distributed_strategy": {
          "default": "ddp",
          "description": "The strategy for distributed training",
          "title": "distributed_strategy",
          "type": "string"
        },
        "gpu_ids": {
          "automl_enabled": false,
          "default": [
            0
          ],
          "description": "\n        List of GPU IDs to run the training on. The length of this list\n        must be equal to the number of gpus in train.num_gpus.",
          "popular": true,
          "title": "GPU IDs",
          "type": "list"
        },
        "is_dry_run": {
          "default": false,
          "description": "Flag to run only one batch for debugging purposes",
          "title": "is_dry_run",
          "type": "bool"
        },
        "loss": {
          "automl_enabled": false,
          "default": {
            "alpha": 5,
            "beta": 10,
            "eps": 1e-06,
            "ohem_ratio": 3,
            "type": "DBLoss"
          },
          "description": "Hyper parameters to configure the loss.",
          "properties": {
            "alpha": {
              "default": 5,
              "description": "The alpha coefficient.",
              "maximum": Infinity,
              "minimum": 1,
              "title": "alpha",
              "type": "int"
            },
            "beta": {
              "default": 10,
              "description": "The beta coefficient",
              "maximum": Infinity,
              "minimum": 1,
              "title": "beta",
              "type": "int"
            },
            "eps": {
              "default": 1e-06,
              "description": "The epsilon coefficient.",
              "maximum": 1.0,
              "minimum": 0.0,
              "title": "epsilon",
              "type": "float"
            },
            "ohem_ratio": {
              "default": 3,
              "description": "The ohem_ratio coefficient",
              "maximum": Infinity,
              "minimum": 1,
              "title": "ohem_ratio",
              "type": "int"
            },
            "type": {
              "default": "DBLoss",
              "description": "Loss function name.",
              "title": "type",
              "type": "string"
            }
          },
          "title": "loss",
          "type": "collection"
        },
        "lr_scheduler": {
          "automl_disabled_parameters": [
            "train.lr_scheduler.args"
          ],
          "automl_enabled": false,
          "default": {
            "args": {
              "warmup_epoch": 3
            },
            "type": "WarmupPolyLR"
          },
          "description": "Hyper parameters to configure the learning rate scheduler.",
          "properties": {
            "args": {
              "automl_enabled": false,
              "default": {
                "warmup_epoch": 3
              },
              "description": "Configurable parameters to construct the learning scheduler.",
              "properties": {
                "warmup_epoch": {
                  "default": 3,
                  "description": "The warmup epoch to the initial learning rate. Should be different from the num_epochs.",
                  "maximum": Infinity,
                  "minimum": 1,
                  "title": "warmup_epoch",
                  "type": "int"
                }
              },
              "type": "collection"
            },
            "type": {
              "default": "WarmupPolyLR",
              "description": "The learning scheduler.",
              "title": "type",
              "type": "string"
            }
          },
          "title": "lr_scheduler",
          "type": "collection"
        },
        "metric": {
          "automl_disabled_parameters": [
            "train.metric.args"
          ],
          "automl_enabled": false,
          "default": {
            "args": {
              "is_output_polygon": false
            },
            "type": "QuadMetric"
          },
          "description": "Hyper parameters to configure the metric.",
          "properties": {
            "args": {
              "automl_enabled": false,
              "default": {
                "is_output_polygon": false
              },
              "description": "Configurable parameters to construct the metric computing.",
              "properties": {
                "is_output_polygon": {
                  "default": false,
                  "description": "Flag to output polygon or BBOX",
                  "title": "is_output_polygon",
                  "type": "bool"
                }
              },
              "type": "collection"
            },
            "type": {
              "default": "QuadMetric",
              "description": "The configuration for metric computing.",
              "title": "type",
              "type": "string"
            }
          },
          "title": "metric",
          "type": "collection"
        },
        "model_ema": {
          "default": false,
          "description": "Flag to enable model EMA",
          "title": "model_ema",
          "type": "bool"
        },
        "model_ema_decay": {
          "default": 0.9999,
          "description": "The decay of model EMA",
          "maximum": 1.0,
          "minimum": 0.0,
          "title": "model_ema_decay",
          "type": "float"
        },
        "num_epochs": {
          "default": 10,
          "description": "Number of epochs to run the training.",
          "maximum": Infinity,
          "minimum": 1,
          "popular": true,
          "title": "Number of epochs",
          "type": "int"
        },
        "num_gpus": {
          "default": 1,
          "description": "The number of GPUs to run the train job.",
          "minimum": 1,
          "popular": true,
          "title": "Number of GPUs",
          "type": "int"
        },
        "num_nodes": {
          "default": 1,
          "description": "Number of nodes to run the training on. If > 1, then multi-node is enabled.",
          "minimum": 1,
          "popular": true,
          "title": "Number of nodes",
          "type": "int"
        },
        "optimizer": {
          "automl_disabled_parameters": [
            "train.optimizer.args"
          ],
          "automl_enabled": false,
          "default": {
            "args": {
              "amsgrad": true,
              "eps": 1e-08,
              "lr": 0.001,
              "momentum": 0.0,
              "weight_decay": 0.0
            },
            "type": "Adam"
          },
          "description": "Hyper parameters to configure the optimizer.",
          "popular": [
            "args"
          ],
          "properties": {
            "args": {
              "automl_default_parameters": [
                "train.optimizer.args.lr",
                "train.optimizer.args.weight_decay"
              ],
              "automl_enabled": false,
              "default": {
                "amsgrad": true,
                "eps": 1e-08,
                "lr": 0.001,
                "momentum": 0.0,
                "weight_decay": 0.0
              },
              "description": "Configurable parameters to construct the optimizer.",
              "popular": [
                "lr"
              ],
              "properties": {
                "amsgrad": {
                  "default": true,
                  "description": "Flag to use AMSGrad as stochastic optimization method",
                  "title": "amsgrad",
                  "type": "bool"
                },
                "eps": {
                  "default": 1e-08,
                  "description": "The epsilon coefficient",
                  "maximum": 1.0,
                  "minimum": 0.0,
                  "title": "epsilon",
                  "type": "float"
                },
                "lr": {
                  "automl_enabled": true,
                  "default": 0.001,
                  "description": "The initial learning rate",
                  "math_cond": "> 0.0",
                  "maximum": Infinity,
                  "minimum": 0.0,
                  "popular": true,
                  "title": "learning rate",
                  "type": "float"
                },
                "momentum": {
                  "default": 0.0,
                  "description": "The momentum for the Adam optimizer.",
                  "maximum": 1.0,
                  "minimum": 0.0,
                  "title": "momentum - Adam",
                  "type": "float"
                },
                "weight_decay": {
                  "automl_enabled": true,
                  "default": 0.0,
                  "description": "The weight decay coefficient.",
                  "math_cond": ">= 0.0",
                  "maximum": Infinity,
                  "minimum": 0.0,
                  "title": "weight decay",
                  "type": "float"
                }
              },
              "type": "collection"
            },
            "type": {
              "default": "Adam",
              "description": "Optimizer type.",
              "title": "type",
              "type": "string"
            }
          },
          "title": "optimizer",
          "type": "collection"
        },
        "post_processing": {
          "automl_disabled_parameters": [
            "train.post_processing.args"
          ],
          "automl_enabled": false,
          "default": {
            "args": {
              "box_thresh": 0.55,
              "max_candidates": 1000,
              "thresh": 0.3,
              "unclip_ratio": 1.5
            },
            "type": "SegDetectorRepresenter"
          },
          "description": "Hyper parameters to configure the post_processing.",
          "popular": [
            "args"
          ],
          "properties": {
            "args": {
              "automl_enabled": false,
              "default": {
                "box_thresh": 0.55,
                "max_candidates": 1000,
                "thresh": 0.3,
                "unclip_ratio": 1.5
              },
              "description": "Configurable parameters to construct the postprocessing.",
              "popular": [
                "thresh",
                "unclip_ratio",
                "box_thresh",
                "max_candidates"
              ],
              "properties": {
                "box_thresh": {
                  "default": 0.55,
                  "description": "The threshold for BBOX.",
                  "maximum": 1.0,
                  "minimum": 0.0,
                  "popular": true,
                  "title": "box_thresh",
                  "type": "float"
                },
                "max_candidates": {
                  "default": 1000,
                  "description": "The maximum candidate BBOX.",
                  "maximum": Infinity,
                  "minimum": 1,
                  "popular": true,
                  "title": "max_candidates",
                  "type": "int"
                },
                "thresh": {
                  "default": 0.3,
                  "description": "The threshold for binarization.",
                  "maximum": 1.0,
                  "minimum": 0.0,
                  "popular": true,
                  "title": "thresh",
                  "type": "float"
                },
                "unclip_ratio": {
                  "default": 1.5,
                  "description": "The unclip ratio using the Vatti clipping algorithm.",
                  "maximum": Infinity,
                  "minimum": 0.0,
                  "popular": true,
                  "title": "unclip_ratio",
                  "type": "float"
                }
              },
              "type": "collection"
            },
            "type": {
              "default": "SegDetectorRepresenter",
              "description": "The postprocessing name.",
              "title": "type",
              "type": "string"
            }
          },
          "title": "post_processing",
          "type": "collection"
        },
        "precision": {
          "default": "fp32",
          "description": "The training precision",
          "title": "precision",
          "type": "string"
        },
        "results_dir": {
          "default": "",
          "description": "\n        Path to where all the assets generated from a task are stored.\n        ",
          "title": "Results directory",
          "type": "string"
        },
        "resume_training_checkpoint_path": {
          "default": "",
          "description": "Path to the checkpoint to resume training from.",
          "title": "Resume checkpoint path",
          "type": "string"
        },
        "seed": {
          "default": 1234,
          "description": "The seed for the initializer in PyTorch. If < 0, disable fixed seed.",
          "maximum": Infinity,
          "minimum": -1,
          "title": "Seed for randomization",
          "type": "int"
        },
        "trainer": {
          "automl_enabled": false,
          "default": {
            "clip_grad_norm": 5.0
          },
          "description": "Hyper parameters to configure the trainer.",
          "properties": {
            "clip_grad_norm": {
              "default": 5.0,
              "description": "\n        Amount to clip the gradient by L2 Norm.\n        A value of 0.0 specifies no clipping.",
              "maximum": Infinity,
              "minimum": 0.0,
              "title": "clip gradient norm",
              "type": "float"
            }
          },
          "title": "trainer",
          "type": "collection"
        },
        "use_distributed_sampler": {
          "default": false,
          "description": "Use distributed sampler for multi-GPU training",
          "title": "use_distributed_sampler",
          "type": "bool"
        },
        "validation_interval": {
          "default": 1,
          "description": "\n        The interval (in epochs) at which a evaluation\n        will be triggered on the validation dataset.",
          "minimum": 1,
          "popular": true,
          "title": "Validation interval",
          "type": "int"
        }
      },
      "type": "collection"
    },
    "wandb": {
      "automl_disabled_parameters": [
        "wandb.tags"
      ],
      "automl_enabled": false,
      "default": {
        "enable": true,
        "entity": "",
        "group": "",
        "name": "TAO Toolkit Training",
        "project": "TAO Toolkit",
        "reinit": false,
        "run_id": "",
        "save_code": false,
        "sync_tensorboard": false,
        "tags": [
          "tao-toolkit"
        ]
      },
      "properties": {
        "enable": {
          "default": true,
          "type": "bool"
        },
        "entity": {
          "default": "",
          "type": "string"
        },
        "group": {
          "default": "",
          "type": "string"
        },
        "name": {
          "default": "TAO Toolkit Training",
          "type": "string"
        },
        "project": {
          "default": "TAO Toolkit",
          "type": "string"
        },
        "reinit": {
          "default": false,
          "type": "bool"
        },
        "run_id": {
          "default": "",
          "type": "string"
        },
        "save_code": {
          "default": false,
          "type": "bool"
        },
        "sync_tensorboard": {
          "default": false,
          "type": "bool"
        },
        "tags": {
          "automl_enabled": false,
          "default": [
            "tao-toolkit"
          ],
          "type": "list"
        }
      },
      "type": "collection"
    }
  },
  "required": [
    "quantize"
  ],
  "type": "object",
  "x_tao_schema": {
    "action": "quantize",
    "core_module": "ocdnet",
    "model": "ocdnet",
    "network_arch": "ocdnet",
    "schema_action": "quantize",
    "schema_version": 1,
    "source": "tao-core dataclass config"
  }
}

Requirements and compatibility

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

Needs: Declares allowed tools, Needs Python

The author says: Requires docker + nvidia-container-toolkit.

Security

What our scanner noticed· Medium risk

Large base64-encoded blobs

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

Created by

NVIDIA Corporation

Publisher

NVIDIA (Official)

Source

NVIDIA/skills · skills/tao-train-ocdnet · commit fce3cca

Licence

Apache License 2.0 (redistribution allowed)

LICENSE (added by AutoGineer from LICENSE-APACHE)
ATTRIBUTION.md (written by AutoGineer)

Synced

2026-11-09

AutoGineer is not affiliated with or endorsed by NVIDIA Corporation or NVIDIA.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download tao-train-ocdnet.zip