Tao Train Segformer

NVIDIAOfficialCreated by NVIDIA Corporation

Skill
Tool integration
Data and AI
Apache-2.0
Medium risk

From the author

SegFormer for semantic segmentation. Lightweight transformer-based architecture with hierarchical feature extraction, efficient for real-time segmentation tasks. Use when training, evaluating, exporting, quantizing, or running inference for a TAO SegFormer model. Trigger phrases include "train SegFormer", "semantic segmentation", "lightweight transformer segmenter", "real-time semantic segmentation".

— the description by NVIDIA Corporation

AutoGineer summary

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

The tao-train-segformer skill provides a comprehensive interface for managing the lifecycle of SegFormer models within the NVIDIA TAO ecosystem. It supports hierarchical feature extraction for efficient semantic segmentation, offering capabilities for training, evaluation, quantization, and export. The skill is designed for real-time applications and includes support for AutoML-driven training to optimize performance metrics like val_miou. It integrates with TAO Deploy workflows for generating and running optimized TensorRT engines.

What it does

Trains SegFormer models with support for AutoML optimization.
Evaluates model performance using test_miou metrics.
Exports trained models to ONNX format for deployment.
Quantizes models to improve inference efficiency.
Provides paths for TensorRT engine generation and inference via TAO Deploy.
Manages checkpointing for training, resuming, and downstream tasks.

When to use it

When you need a lightweight, real-time semantic segmentation model.
When you want to leverage transformer-based architectures for image segmentation.
When you are working within the NVIDIA TAO Toolkit environment.
When you need to optimize segmentation models for production deployment on NVIDIA hardware.

Advantages

Efficient, hierarchical transformer architecture optimized for real-time performance.
Seamless integration with TAO AutoML for automated hyperparameter tuning.
Supports standard TAO workflows including quantization and TensorRT deployment.
Flexible configuration for various backbone types and input resolutions.

Usage examples

“Train a SegFormer model on my segmentation dataset.”: The agent initiates the training process using the specified dataset root and default AutoML settings to maximize val_miou.
“Export my trained SegFormer model to ONNX.”: The agent performs the export action, generating an ONNX file based on the provided checkpoint and input dimensions.
“Quantize my SegFormer model for better inference speed.”: The agent runs the quantization process using the provided calibration dataset and model checkpoint.
“Evaluate my SegFormer model performance.”: The agent runs the evaluation action against the test dataset to report the final test_miou metric.

Install

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

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

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

Documentation by NVIDIA Corporation

SegFormer

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).

SegFormer for semantic segmentation. Lightweight transformer-based architecture with hierarchical feature extraction. Efficient for real-time segmentation tasks.

Set model.backbone.pretrained_backbone_path for backbone weights.

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

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.

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.

Supported Actions

The packaged SegFormer PyT CLI supports train, evaluate, export, inference, quantize, and default_specs. This model skill exposes train, evaluate, export, inference, and quantize; resume/retrain is performed through train with train.resume_training_checkpoint_path.

The parent PyT CLI does not expose gen_trt_engine. Use models/segformer/deploy for TensorRT engine generation, TensorRT evaluation, and TensorRT inference.

Training Requirements

Dataset type: segmentation
Formats: unet
AutoML training metric: val_miou (maximize).
Standalone evaluation metric: test_miou. Use the train-status val_miou for AutoML ranking and the evaluator's test_miou only for checkpoint verification.

Per-Action Dataset Requirements

Action	Spec Key	Source	Files	List?
evaluate	dataset.segment.root_dir	eval_dataset	extracted root containing `images/<split>` and `masks/<split>`	No
export	dataset.segment.root_dir	train_datasets	extracted root containing `images/<split>` and `masks/<split>`	No
inference	dataset.segment.root_dir	inference_dataset	extracted root containing `images/<split>` and `masks/<split>`	No
quantize	dataset.segment.root_dir	train_datasets	extracted root containing `images/<split>` and `masks/<split>`	No
quantize	dataset.segment.quant_calibration_dataset.images_dir	calibration_dataset	extracted image directory	No
train	dataset.segment.root_dir	train_datasets	extracted root containing `images/<split>` and `masks/<split>`	No

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.

SEG_TRAIN_ROOT = "/data/segformer/train"
SEG_EVAL_ROOT = "/data/segformer/eval"
SEG_INFER_ROOT = "/data/segformer/infer"
CAL_IMAGES = f"{SEG_TRAIN_ROOT}/images/train"

train (mandatory data sources):

{
    "train.num_gpus": 1,
    "train.num_epochs": 10,
    "train.checkpoint_interval": 10,
    "train.validation_interval": 10,
    "dataset.segment.batch_size": 4,
    "dataset.segment.root_dir": SEG_TRAIN_ROOT,
}

evaluate (mandatory data sources):

{
    "evaluate.batch_size": 4,
    "dataset.segment.root_dir": SEG_EVAL_ROOT,
    "evaluate.checkpoint": CHECKPOINT,
}

inference (mandatory data sources):

{
    "dataset.segment.batch_size": 1,
    "dataset.segment.root_dir": SEG_INFER_ROOT,
    "inference.checkpoint": CHECKPOINT,
}

export (mandatory data sources):

{
    "dataset.segment.root_dir": SEG_TRAIN_ROOT,
    "export.checkpoint": CHECKPOINT,
    "export.input_height": 256,
    "export.input_width": 256,
    "export.onnx_file": ONNX_FILE,
}

quantize (mandatory data sources):

{
    "dataset.segment.root_dir": SEG_TRAIN_ROOT,
    "dataset.segment.quant_calibration_dataset.images_dir": CAL_IMAGES,
    "quantize.model_path": CHECKPOINT,
}

If the source dataset is delivered as separate images/*.tar.gz and masks/*.tar.gz archives, extract them before launch so root_dir contains directories such as images/train, images/val, images/test, masks/train, and masks/val. Do not point dataset.segment.root_dir at an archive staging folder that still contains only tarballs.

Eval Dataset

Optional. Validation data is typically part of the root_dir structure.

Important Parameters

dataset.segment.num_classes: Number of segmentation classes. Default 2 (binary). Must match the number of classes in your mask annotations.
model.backbone.type: Default fan_small_12_p4_hybrid. Supported includes FAN variants, SegFormer MIT variants, and others.
dataset.segment.root_dir: Root directory of the segmentation dataset.
dataset.segment.img_size: Input image size. Default 256. Increase for finer segmentation at the cost of memory.
train.optim.lr: Learning rate. Default 6e-5.
model.freeze_backbone: Whether to freeze the backbone during training. Useful for fine-tuning with limited data.
dataset.segment.batch_size: Per-GPU batch size. Default 8.
dataset.segment.label_transform: Use the string "None" when no label transform is desired. Do not set this to JSON/YAML null; strict schema merge treats the field as a string enum.
dataset.segment.palette: For grayscale masks, use one integer per RGB entry, for example rgb: [85]. Preserve the dataset's actual label ids and class names rather than normalizing them unless the user explicitly asks for a conversion.

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.num_nodes`	Number of nodes	1
`train.sync_batchnorm`	Sync BN across GPUs	configurable
`train.use_distributed_sampler`	Use distributed sampler	configurable

Multi-GPU strategy: ddp_find_unused_parameters_true
No fsdp support

Multi-node env vars (set by orchestrator): WORLD_SIZE, NODE_RANK, MASTER_ADDR, MASTER_PORT, NUM_GPU_PER_NODE.

Hardware

Minimum 1 GPU(s), recommended 2 GPU(s). 16GB+ (V100 or A100) VRAM per GPU. SegFormer is relatively lightweight. Default img_size=256 is memory-friendly. Increase img_size for higher resolution at the cost of memory and speed.

Error Patterns

CUDA out of memory: Reduce batch_size or img_size. SegFormer memory scales quadratically with image size.

num_classes mismatch: Ensure dataset.segment.num_classes matches the actual number of classes in your mask annotations.

TensorBoard unsupported for segmentation training: Keep train.tensorboard.enabled: false. The SegFormer training entrypoint asserts that TensorBoard visualization is not supported for segmentation, so do not enable TensorBoard just to extract AutoML metrics; use log parsing or a post-train evaluator instead.

AutoML metric extraction: SegFormer train status files report val_miou alongside val_loss, val_acc, and other validation KPIs. Default AutoML train launches must optimize val_miou with direction: maximize; do not optimize val_loss for default model invocations.

For AutoML or long segmentation sweeps, read val_miou from results_dir/train/status.json first. If the wrapper reports a terminal failure but the structured status file reached the configured training budget and contains finite val_miou, report the recovered metric with the wrapper failure noted instead of discarding the measurement.

For high-resolution custom segmentation targets, keep dataset paths as per-run inputs. Do not add customer/user-specific roots to this reusable skill. When the user asks for a fixed full-budget search, remember that bracket algorithms (asha, bohb, dehb, hyperband, hyperband_es, pbt) may intentionally lower train.num_epochs for some recommendations; use Bayesian/BFBO or lock the budget if every recommendation must run the full epoch count.

Checkpoint handoff: For evaluate/export/inference/quantize/resume, use the checkpoint resolver on the best AutoML child job's results_dir/train/ folder and select the action-appropriate model_epoch_*.pth checkpoint, such as model_epoch_000_step_00010.pth. SegFormer may also write segformer_model_latest.pth, but that should only be used when a caller explicitly requests latest. Preserve dataset.segment.num_classes, dataset.segment.img_size, and dataset.segment.root_dir overrides for downstream actions.

Resume/retrain checkpoint: Resume uses train.resume_training_checkpoint_path. Pass the exact resolved checkpoint from the previous train output, not a guessed model.pth path. A resumed one-epoch run should produce the next checkpoint in the new results directory, for example model_epoch_001_step_00020.pth.

Export / TensorRT shape alignment: Keep export.input_height and export.input_width aligned with dataset.segment.img_size unless the trained model and deploy specs have been validated at another resolution. The packaged fresh-install path is validated at 256x256, matching the default SegFormer dataset and deploy templates.

Parent segformer gen_trt_engine rejected by the PyT CLI: In the validated 7.0.0 PyT container, segformer gen_trt_engine is not a valid parent-model subtask. Use the SegFormer deploy workflow (references/tao-deploy-segformer.md) for TensorRT engine generation, TensorRT evaluation, and TensorRT inference.

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.

Inference mappings from TAO Core segformer.config.json:

Action	Spec Field	Inference Function	Meaning
evaluate	`encryption_key`	`key`	encryption key
evaluate	`evaluate.checkpoint`	`parent_model`	model file inferred from the parent job results folder
evaluate	`evaluate.trt_engine`	`parent_model`	model file inferred from the parent job results folder
evaluate	`results_dir`	`output_dir`	current job results directory
export	`encryption_key`	`key`	encryption key
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	`encryption_key`	`key`	encryption key
inference	`inference.checkpoint`	`parent_model`	model file inferred from the parent job results folder
inference	`inference.trt_engine`	`parent_model`	model file inferred from the parent job results folder
inference	`results_dir`	`output_dir`	current job results directory
quantize	`encryption_key`	`key`	encryption key
quantize	`quantize.model_path`	`parent_model`	model file inferred from the parent job results folder
quantize	`results_dir`	`output_dir`	current job results directory
train	`encryption_key`	`key`	encryption key
train	`model.backbone.pretrained_backbone_path`	`ptm_if_no_resume_model`	PTM when no resume checkpoint exists
train	`results_dir`	`output_dir`	current job results directory
train	`train.pretrained_model_path`	`ptm_if_no_resume_model`	PTM when no resume checkpoint exists
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-segformer

Files

ATTRIBUTION.mdAdded by AutoGineer927 B
BENCHMARK.md6.9 KiB
LICENSELicence file added by AutoGineer11.2 KiB
SKILL.md14.5 KiB
config/skillspector-baseline.yaml834 B
evals/evals.json776 B
references/skill_info.yaml4.4 KiB
references/spec_template_deploy_evaluate.yaml748 B
references/spec_template_deploy_gen_trt_engine.yaml1.4 KiB
references/spec_template_deploy_inference.yaml749 B
references/spec_template_evaluate.yaml2.8 KiB
references/spec_template_export.yaml2.9 KiB
references/spec_template_inference.yaml2.8 KiB
references/spec_template_quantize.yaml2.6 KiB
references/spec_template_train.yaml2.6 KiB
references/tao-deploy-segformer.md6.0 KiB
references/tao-deploy-segformer.skill_info.yaml2.8 KiB
schemas/evaluate.schema.json48.4 KiB
schemas/export.schema.json49.3 KiB
schemas/inference.schema.json48.4 KiB
schemas/manifest.json17.7 KiB
schemas/quantize.schema.json45.6 KiB
schemas/train.schema.json45.6 KiB
skill-card.md4.0 KiB
skill.oms.sig8.9 KiB

schemas/evaluate.schema.json

{
  "automl_default_parameters": [
    "dataset.segment.augmentation.random_color.contrast",
    "dataset.segment.augmentation.random_color.saturation",
    "dataset.segment.augmentation.with_scale_random_crop.enable",
    "train.optim.weight_decay",
    "dataset.segment.augmentation.random_rotate.rotate_probability",
    "dataset.segment.augmentation.random_color.brightness",
    "dataset.segment.augmentation.random_color.hue",
    "train.optim.momentum",
    "model.backbone.freeze_backbone",
    "dataset.segment.augmentation.random_flip.hflip_probability",
    "dataset.segment.augmentation.random_color.color_probability",
    "dataset.segment.augmentation.random_rotate.enable",
    "train.optim.lr",
    "dataset.segment.augmentation.random_flip.enable",
    "dataset.segment.augmentation.random_color.enable",
    "dataset.segment.augmentation.random_flip.vflip_probability"
  ],
  "automl_disabled_parameters": [
    "dataset.segment.augmentation.std",
    "train.cudnn",
    "quantize",
    "dataset.segment.quant_calibration_dataset",
    "gen_trt_engine.tensorrt.calibration.cal_image_dir",
    "dataset.segment.augmentation",
    "quantize.backend_kwargs",
    "train.gpu_ids",
    "model.decode_head.feature_strides",
    "wandb.tags",
    "model.backbone",
    "quantize.skip_names",
    "dataset.segment.palette",
    "train.tensorboard",
    "dataset.segment.augmentation.random_rotate",
    "evaluate",
    "inference",
    "train",
    "gen_trt_engine",
    "gen_trt_engine.tensorrt.layers_precision",
    "model.decode_head.in_channels",
    "dataset",
    "dataset.segment",
    "gen_trt_engine.tensorrt",
    "quantize.layers",
    "dataset.segment.augmentation.random_flip",
    "train.segment",
    "model.decode_head",
    "model",
    "dataset.segment.augmentation.with_scale_random_crop.scale_range",
    "dataset.segment.augmentation.mean",
    "evaluate.gpu_ids",
    "train.optim",
    "gen_trt_engine.tensorrt.calibration",
    "train.segment.weights",
    "dataset.segment.augmentation.random_rotate.angle_list",
    "export",
    "dataset.segment.augmentation.with_scale_random_crop",
    "wandb",
    "dataset.segment.augmentation.random_color",
    "inference.gpu_ids",
    "model.decode_head.decoder_params",
    "model.decode_head.in_index"
  ],
  "default": {
    "dataset": {
      "segment": {
        "augmentation": {
          "mean": [
            0.5,
            0.5,
            0.5
          ],
          "random_color": {
            "brightness": 0.3,
            "color_probability": 0.5,
            "contrast": 0.3,
            "enable": true,
            "hue": 0.3,
            "saturation": 0.3
          },
          "random_flip": {
            "enable": true,
            "hflip_probability": 0.5,
            "vflip_probability": 0.5
          },
          "random_rotate": {
            "angle_list": [
              90,
              180,
              270
            ],
            "enable": true,
            "rotate_probability": 0.5
          },
          "std": [
            0.5,
            0.5,
            0.5
          ],
          "with_random_blur": true,
          "with_random_crop": true,
          "with_scale_random_crop": {
            "enable": true,
            "scale_range": [
              1,
              1.2
            ]
          }
        },
        "batch_size": 8,
        "dataset": "SFDataset",
        "img_size": 256,
        "label_transform": "norm",
        "num_classes": 2,
        "palette": [
          {
            "label_id": 0,
            "mapping_class": "foreground",
            "rgb": [
              0,
              0,
              0
            ],
            "seg_class": "foreground"
          },
          {
            "label_id": 1,
            "mapping_class": "background",
            "rgb": [
              1,
              1,
              1
            ],
            "seg_class": "background"
          }
        ],
        "predict_split": "test",
        "quant_calibration_dataset": {
          "images_dir": ""
        },
        "root_dir": "???",
        "shuffle": true,
        "test_split": "val",
        "train_split": "train",
        "validation_split": "val",
        "workers": 1
      }
    },
    "encryption_key": "",
    "evaluate": {
      "batch_size": 8,
      "checkpoint": "???",
      "gpu_ids": [
        0
      ],
      "num_gpus": 1,
      "num_nodes": 1,
      "results_dir": "",
      "trt_engine": "",
      "vis_after_n_batches": 1
    },
    "model": {
      "backbone": {
        "feat_downsample": false,
        "freeze_backbone": false,
        "pretrained_backbone_path": "",
        "type": "fan_small_12_p4_hybrid"
      },
      "decode_head": {
        "align_corners": false,
        "decoder_params": {
          "embed_dim": 768
        },
        "feature_strides": [
          4,
          8,
          16,
          32
        ],
        "in_channels": [
          64,
          128,
          320,
          512
        ],
        "in_index": [
          0,
          1,
          2,
          3
        ]
      }
    },
    "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
      },
      "gpu_ids": [
        0
      ],
      "num_epochs": 10,
      "num_gpus": 1,
      "num_nodes": 1,
      "optim": {
        "lr": 6e-05,
        "momentum": 0.9,
        "monitor_name": "val_loss",
        "optim": "adamw",
        "policy": "linear",
        "weight_decay": 0.01
      },
      "pretrained_model_path": "",
      "results_dir": "",
      "resume_training_checkpoint_path": "",
      "seed": 1234,
      "segment": {
        "loss": "ce",
        "weights": [
          0.5,
          0.5,
          0.5,
          0.8,
          1.0
        ]
      },
      "sync_batchnorm": false,
      "tensorboard": {
        "enabled": false,
        "infrequent_logging_frequency": 2
      },
      "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": {
    "evaluate": {
      "gpu_ids": [
        0
      ],
      "num_gpus": 1,
      "num_nodes": 1
    },
    "gen_trt_engine": {
      "batch_size": -1,
      "gpu_id": 0,
      "tensorrt": {
        "calibration": {
          "cal_batch_size": 1,
          "cal_batches": 1
        },
        "max_batch_size": 1,
        "min_batch_size": 1,
        "opt_batch_size": 1
      }
    },
    "inference": {
      "gpu_ids": [
        0
      ],
      "num_gpus": 1,
      "num_nodes": 1
    },
    "train": {
      "checkpoint_interval": 1,
      "gpu_ids": [
        0
      ],
      "num_epochs": 10,
      "num_gpus": 1,
      "num_nodes": 1,
      "validation_interval": 1
    }
  },
  "properties": {
    "automl_disabled_parameters": [
      "wandb",
      "model",
      "dataset",
      "train",
      "evaluate",
      "inference",
      "export",
      "gen_trt_engine",
      "quantize"
    ],
    "dataset": {
      "automl_disabled_parameters": [
        "dataset.segment"
      ],
      "automl_enabled": false,
      "default": {
        "segment": {
          "augmentation": {
            "mean": [
              0.5,
              0.5,
              0.5
            ],
            "random_color": {
              "brightness": 0.3,
              "color_probability": 0.5,
              "contrast": 0.3,
              "enable": true,
              "hue": 0.3,
              "saturation": 0.3
            },
            "random_flip": {
              "enable": true,
              "hflip_probability": 0.5,
              "vflip_probability": 0.5
            },
            "random_rotate": {
              "angle_list": [
                90,
                180,
                270
              ],
              "enable": true,
              "rotate_probability": 0.5
            },
            "std": [
              0.5,
              0.5,
              0.5
            ],
            "with_random_blur": true,
            "with_random_crop": true,
            "with_scale_random_crop": {
              "enable": true,
              "scale_range": [
                1,
                1.2
              ]
            }
          },
          "batch_size": 8,
          "dataset": "SFDataset",
          "img_size": 256,
          "label_transform": "norm",
          "num_classes": 2,
          "palette": [
            {
              "label_id": 0,
              "mapping_class": "foreground",
              "rgb": [
                0,
                0,
                0
              ],
              "seg_class": "foreground"
            },
            {
              "label_id": 1,
              "mapping_class": "background",
              "rgb": [
                1,
                1,
                1
              ],
              "seg_class": "background"
            }
          ],
          "predict_split": "test",
          "quant_calibration_dataset": {
            "images_dir": ""
          },
          "root_dir": "???",
          "shuffle": true,
          "test_split": "val",
          "train_split": "train",
          "validation_split": "val",
          "workers": 1
        }
      },
      "properties": {
        "segment": {
          "automl_disabled_parameters": [
            "dataset.segment.augmentation",
            "dataset.segment.palette",
            "dataset.segment.quant_calibration_dataset"
          ],
          "automl_enabled": false,
          "default": {
            "augmentation": {
              "mean": [
                0.5,
                0.5,
                0.5
              ],
              "random_color": {
                "brightness": 0.3,
                "color_probability": 0.5,
                "contrast": 0.3,
                "enable": true,
                "hue": 0.3,
                "saturation": 0.3
              },
              "random_flip": {
                "enable": true,
                "hflip_probability": 0.5,
                "vflip_probability": 0.5
              },
              "random_rotate": {
                "angle_list": [
                  90,
                  180,
                  270
                ],
                "enable": true,
                "rotate_probability": 0.5
              },
              "std": [
                0.5,
                0.5,
                0.5
              ],
              "with_random_blur": true,
              "with_random_crop": true,
              "with_scale_random_crop": {
                "enable": true,
                "scale_range": [
                  1,
                  1.2
                ]
              }
            },
            "batch_size": 8,
            "dataset": "SFDataset",
            "img_size": 256,
            "label_transform": "norm",
            "num_classes": 2,
            "palette": [
              {
                "label_id": 0,
                "mapping_class": "foreground",
                "rgb": [
                  0,
                  0,
                  0
                ],
                "seg_class": "foreground"
              },
              {
                "label_id": 1,
                "mapping_class": "background",
                "rgb": [
                  1,
                  1,
                  1
                ],
                "seg_class": "background"
              }
            ],
            "predict_split": "test",
            "quant_calibration_dataset": {
              "images_dir": ""
            },
            "root_dir": "???",
            "shuffle": true,
            "test_split": "val",
            "train_split": "train",
            "validation_split": "val",
            "workers": 1
          },
          "properties": {
            "augmentation": {
              "automl_disabled_parameters": [
                "dataset.segment.augmentation.random_flip",
                "dataset.segment.augmentation.random_rotate",
                "dataset.segment.augmentation.random_color",
                "dataset.segment.augmentation.with_scale_random_crop",
                "dataset.segment.augmentation.mean",
                "dataset.segment.augmentation.std"
              ],
              "automl_enabled": false,
              "default": {
                "mean": [
                  0.5,
                  0.5,
                  0.5
                ],
                "random_color": {
                  "brightness": 0.3,
                  "color_probability": 0.5,
                  "contrast": 0.3,
                  "enable": true,
                  "hue": 0.3,
                  "saturation": 0.3
                },
                "random_flip": {
                  "enable": true,
                  "hflip_probability": 0.5,
                  "vflip_probability": 0.5
                },
                "random_rotate": {
                  "angle_list": [
                    90,
                    180,
                    270
                  ],
                  "enable": true,
                  "rotate_probability": 0.5
                },
                "std": [
                  0.5,
                  0.5,
                  0.5
                ],
                "with_random_blur": true,
                "with_random_crop": true,
                "with_scale_random_crop": {
                  "enable": true,
                  "scale_range": [
                    1,
                    1.2
                  ]
                }
              },
              "properties": {
                "mean": {
                  "automl_enabled": false,
                  "default": [
                    0.5,
                    0.5,
                    0.5
                  ],
                  "description": "Mean for the augmentation",
                  "title": "Mean",
                  "type": "list"
                },
                "random_color": {
                  "automl_default_parameters": [
                    "dataset.segment.augmentation.random_color.brightness",
                    "dataset.segment.augmentation.random_color.contrast",
                    "dataset.segment.augmentation.random_color.saturation",
                    "dataset.segment.augmentation.random_color.hue",
                    "dataset.segment.augmentation.random_color.enable",
                    "dataset.segment.augmentation.random_color.color_probability"
                  ],
                  "automl_enabled": false,
                  "default": {
                    "brightness": 0.3,
                    "color_probability": 0.5,
                    "contrast": 0.3,
                    "enable": true,
                    "hue": 0.3,
                    "saturation": 0.3
                  },
                  "properties": {
                    "brightness": {
                      "automl_enabled": true,
                      "default": 0.3,
                      "description": "Random Color Brightness (torchvision ColorJitter range)",
                      "maximum": 2.0,
                      "minimum": 0.0,
                      "type": "float"
                    },
                    "color_probability": {
                      "automl_enabled": true,
                      "default": 0.5,
                      "description": "Random Color Probability",
                      "maximum": 1.0,
                      "minimum": 0.0,
                      "type": "float"
                    },
                    "contrast": {
                      "automl_enabled": true,
                      "default": 0.3,
                      "description": "Random Color Contrast (torchvision ColorJitter range)",
                      "maximum": 2.0,
                      "minimum": 0.0,
                      "type": "float"
                    },
                    "enable": {
                      "automl_enabled": true,
                      "default": true,
                      "description": "Flag to enable Random Color",
                      "type": "bool"
                    },
                    "hue": {
                      "automl_enabled": true,
                      "default": 0.3,
                      "description": "Random Color Hue (torchvision ColorJitter requires 0 <= hue <= 0.5)",
                      "maximum": 0.5,
                      "minimum": 0.0,
                      "type": "float"
                    },
                    "saturation": {
                      "automl_enabled": true,
                      "default": 0.3,
                      "description": "Random Color Saturation (torchvision ColorJitter range)",
                      "maximum": 2.0,
                      "minimum": 0.0,
                      "type": "float"
                    }
                  },
                  "type": "collection"
                },
                "random_flip": {
                  "automl_default_parameters": [
                    "dataset.segment.augmentation.random_flip.vflip_probability",
                    "dataset.segment.augmentation.random_flip.hflip_probability",
                    "dataset.segment.augmentation.random_flip.enable"
                  ],
                  "automl_enabled": false,
                  "default": {
                    "enable": true,
                    "hflip_probability": 0.5,
                    "vflip_probability": 0.5
                  },
                  "properties": {
                    "enable": {
                      "automl_enabled": true,
                      "default": true,
                      "description": "Flag to enable augmentation",
                      "type": "bool"
                    },
                    "hflip_probability": {
                      "automl_enabled": true,
                      "default": 0.5,
                      "description": "Horizontal Flip probability",
                      "maximum": 1.0,
                      "minimum": 0.0,
                      "type": "float"
                    },
                    "vflip_probability": {
                      "automl_enabled": true,
                      "default": 0.5,
                      "description": "Vertical Flip probability",
                      "maximum": 1.0,
                      "minimum": 0.0,
                      "type": "float"
                    }
                  },
                  "type": "collection"
                },
                "random_rotate": {
                  "automl_default_parameters": [
                    "dataset.segment.augmentation.random_rotate.rotate_probability",
                    "dataset.segment.augmentation.random_rotate.enable"
                  ],
                  "automl_disabled_parameters": [
                    "dataset.segment.augmentation.random_rotate.angle_list"
                  ],
                  "automl_enabled": false,
                  "default": {
                    "angle_list": [
                      90,
                      180,
                      270
                    ],
                    "enable": true,
                    "rotate_probability": 0.5
                  },
                  "properties": {
                    "angle_list": {
                      "automl_enabled": false,
                      "default": [
                        90,
                        180,
                        270
                      ],
                      "description": "Random rotate angle probability",
                      "type": "list"
                    },
                    "enable": {
                      "automl_enabled": true,
                      "default": true,
                      "description": "Flag to enable augmentation",
                      "type": "bool"
                    },
                    "rotate_probability": {
                      "automl_enabled": true,
                      "default": 0.5,
                      "description": "Random Rotate probability",
                      "maximum": 1.0,
                      "minimum": 0.0,
                      "type": "float"
                    }
                  },
                  "type": "collection"
                },
                "std": {
                  "automl_enabled": false,
                  "default": [
                    0.5,
                    0.5,
                    0.5
                  ],
                  "description": "Standard deviation for the augmentation",
                  "title": "Standard Deviation",
                  "type": "list"
                },
                "with_random_blur": {
                  "default": true,
                  "description": "Flag to enable with_random_blur",
                  "type": "bool"
                },
                "with_random_crop": {
                  "default": true,
                  "description": "Flag to enable with_random_crop",
                  "type": "bool"
                },
                "with_scale_random_crop": {
                  "automl_default_parameters": [
                    "dataset.segment.augmentation.with_scale_random_crop.enable"
                  ],
                  "automl_disabled_parameters": [
                    "dataset.segment.augmentation.with_scale_random_crop.scale_range"
                  ],
                  "automl_enabled": false,
                  "default": {
                    "enable": true,
                    "scale_range": [
                      1,
                      1.2
                    ]
                  },
                  "properties": {
                    "enable": {
                      "automl_enabled": true,
                      "default": true,
                      "description": "Flag to enable Random Crop with Scale",
                      "type": "bool"
                    },
                    "scale_range": {
                      "automl_enabled": false,
                      "default": [
                        1,
                        1.2
                      ],
                      "description": "Random Scale range",
                      "type": "list"
                    }
                  },
                  "type": "collection"
                }
              },
              "type": "collection"
            },
            "batch_size": {
              "default": 8,
              "description": "Batch size",
              "maximum": Infinity,
              "minimum": 1,
              "title": "Batch Size",
              "type": "int"
            },
            "dataset": {
              "default": "SFDataset",
              "description": "dataset class",
              "enum": [
                "SFDataset"
              ],
              "type": "categorical"
            },
            "img_size": {
              "default": 256,
              "description": "The input image size",
              "type": "int"
            },
            "label_transform": {
              "default": "norm",
              "description": "label transform",
              "enum": [
                "norm",
                "None"
              ],
              "type": "categorical"
            },
            "num_classes": {
              "default": 2,
              "description": "The number of classes in the training data",
              "math_cond": ">0",
              "maximum": Infinity,
              "minimum": 2,
              "type": "int"
            },
            "palette": {
              "automl_enabled": false,
              "default": [
                {
                  "label_id": 0,
                  "mapping_class": "foreground",
                  "rgb": [
                    0,
                    0,
                    0
                  ],
                  "seg_class": "foreground"
                },
                {
                  "label_id": 1,
                  "mapping_class": "background",
                  "rgb": [
                    1,
                    1,
                    1
                  ],
                  "seg_class": "background"
                }
              ],
              "description": "Palette, be careful of label_transform, if norm then RGB value from 0~1, else 0~255",
              "title": "Palette",
              "type": "list"
            },
            "predict_split": {
              "default": "test",
              "description": "Predict split folder name",
              "type": "string"
            },
            "quant_calibration_dataset": {
              "automl_enabled": false,
              "default": {
                "images_dir": ""
              },
              "description": "Configurable parameters for the quantization calibration dataset.",
              "properties": {
                "images_dir": {
                  "default": "",
                  "description": "Path to images directory for quantization calibration",
                  "title": "images directory",
                  "type": "string"
                }
              },
              "type": "collection"
            },
            "root_dir": {
              "default": "???",
              "description": "Path to root directory for dataset",
              "type": "string"
            },
            "shuffle": {
              "default": true,
              "description": "Shuffle dataloader",
              "type": "bool"
            },
            "test_split": {
              "default": "val",
              "description": "Test split folder name",
              "type": "string"
            },
            "train_split": {
              "default": "train",
              "description": "Train split folder name",
              "type": "string"
            },
            "validation_split": {
              "default": "val",
              "description": "Validation split folder name",
              "type": "string"
            },
            "workers": {
              "default": 1,
              "description": "Workers",
              "maximum": Infinity,
              "minimum": 0,
              "title": "Workers",
              "type": "int"
            }
          },
          "type": "collection"
        }
      },
      "type": "collection"
    },
    "encryption_key": {
      "default": "",
      "description": "Key for encrypting model checkpoints",
      "title": "Encryption key",
      "type": "string"
    },
    "evaluate": {
      "automl_disabled_parameters": [
        "evaluate.gpu_ids"
      ],
      "automl_enabled": false,
      "default": {
        "batch_size": 8,
        "checkpoint": "???",
        "gpu_ids": [
          0
        ],
        "num_gpus": 1,
        "num_nodes": 1,
        "results_dir": "",
        "trt_engine": "",
        "vis_after_n_batches": 1
      },
      "popular": [
        "num_gpus",
        "num_nodes",
        "gpu_ids"
      ],
      "properties": {
        "batch_size": {
          "default": 8,
          "description": "Batch size",
          "maximum": Infinity,
          "minimum": 1,
          "title": "Batch Size",
          "type": "int"
        },
        "checkpoint": {
          "default": "???",
          "description": "Path to checkpoint file",
          "title": "Path to checkpoint file",
          "type": "string"
        },
        "gpu_ids": {
          "automl_enabled": false,
          "default": [
            0
          ],
          "description": "\n        List of GPU IDs to run the evaluation on. The length of this list\n        must be equal to the number of gpus in evaluate.num_gpus.",
          "popular": true,
          "title": "GPU IDs",
          "type": "list"
        },
        "num_gpus": {
          "default": 1,
          "description": "The number of GPUs to run the evaluation job.",
          "minimum": 1,
          "popular": true,
          "title": "Number of GPUs",
          "type": "int"
        },
        "num_nodes": {
          "default": 1,
          "description": "Number of nodes to run the evaluation on. If > 1, then multi-node is enabled.",
          "minimum": 1,
          "popular": true,
          "title": "Number of nodes",
          "type": "int"
        },
        "results_dir": {
          "default": "",
          "description": "\n        Path to where all the assets generated from a task are stored.\n        ",
          "title": "Results directory",
          "type": "string"
        },
        "trt_engine": {
          "default": "",
          "description": "Path to the TensorRT engine to be used for evaluation.\n                    This only works with :code:`tao-deploy`.",
          "title": "TensorRT Engine",
          "type": "string"
        },
        "vis_after_n_batches": {
          "default": 1,
          "description": "Visualize evaluation segmentation results after n batches",
          "maximum": Infinity,
          "minimum": 1,
          "type": "int"
        }
      },
      "type": "collection"
    },
    "model": {
      "automl_disabled_parameters": [
        "model.backbone",
        "model.decode_head"
      ],
      "automl_enabled": false,
      "default": {
        "backbone": {
          "feat_downsample": false,
          "freeze_backbone": false,
          "pretrained_backbone_path": "",
          "type": "fan_small_12_p4_hybrid"
        },
        "decode_head": {
          "align_corners": false,
          "decoder_params": {
            "embed_dim": 768
          },
          "feature_strides": [
            4,
            8,
            16,
            32
          ],
          "in_channels": [
            64,
            128,
            320,
            512
          ],
          "in_index": [
            0,
            1,
            2,
            3
          ]
        }
      },
      "properties": {
        "backbone": {
          "automl_default_parameters": [
            "model.backbone.freeze_backbone"
          ],
          "automl_enabled": false,
          "default": {
            "feat_downsample": false,
            "freeze_backbone": false,
            "pretrained_backbone_path": "",
            "type": "fan_small_12_p4_hybrid"
          },
          "properties": {
            "feat_downsample": {
              "default": false,
              "description": "Feature downsample for fan base backbone",
              "title": "Feature downsample",
              "type": "bool"
            },
            "freeze_backbone": {
              "automl_enabled": true,
              "default": false,
              "description": "Flag to freeze backbone",
              "type": "bool"
            },
            "pretrained_backbone_path": {
              "default": "",
              "description": "Path to the pretrained model",
              "type": "string"
            },
            "type": {
              "default": "fan_small_12_p4_hybrid",
              "description": "Backbone architure",
              "enum": [
                "fan_tiny_8_p4_hybrid",
                "fan_large_16_p4_hybrid",
                "fan_small_12_p4_hybrid",
                "fan_base_16_p4_hybrid",
                "vit_large_nvdinov2",
                "vit_giant_nvdinov2",
                "vit_small_dinov3",
                "vit_small_plus_dinov3",
                "vit_base_dinov3",
                "vit_large_dinov3",
                "vit_huge_plus_dinov3",
                "vit_base_nvclip_16_siglip",
                "vit_huge_nvclip_14_siglip"
              ],
              "title": "Backbone architectures",
              "type": "categorical"
            }
          },
          "type": "collection"
        },
        "decode_head": {
          "automl_disabled_parameters": [
            "model.decode_head.in_channels",
            "model.decode_head.in_index",
            "model.decode_head.feature_strides",
            "model.decode_head.decoder_params"
          ],
          "automl_enabled": false,
          "default": {
            "align_corners": false,
            "decoder_params": {
              "embed_dim": 768
            },
            "feature_strides": [
              4,
              8,
              16,
              32
            ],
            "in_channels": [
              64,
              128,
              320,
              512
            ],
            "in_index": [
              0,
              1,
              2,
              3
            ]
          },
          "properties": {
            "align_corners": {
              "default": false,
              "description": "Align corners for the head",
              "title": "Align Corners",
              "type": "bool"
            },
            "decoder_params": {
              "automl_enabled": false,
              "default": {
                "embed_dim": 768
              },
              "description": "Decoder parameters for the head",
              "title": "Decoder Parameters",
              "type": "collection"
            },
            "feature_strides": {
              "automl_enabled": false,
              "default": [
                4,
                8,
                16,
                32
              ],
              "description": "Feature strides for the head",
              "title": "Feature Strides",
              "type": "list"
            },
            "in_channels": {
              "automl_enabled": false,
              "default": [
                64,
                128,
                320,
                512
              ],
              "description": "number of input channels to decoder",
              "type": "list"
            },
            "in_index": {
              "automl_enabled": false,
              "default": [
                0,
                1,
                2,
                3
              ],
              "description": "Input index for the head",
              "title": "Input Index",
              "type": "list"
            }
          },
          "type": "collection"
        }
      },
      "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 to run model quantization for a SegFormer experiment.",
      "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.optim",
        "train.segment",
        "train.tensorboard"
      ],
      "automl_enabled": false,
      "default": {
        "checkpoint_interval": 1,
        "checkpoint_interval_unit": "epoch",
        "cudnn": {
          "benchmark": false,
          "deterministic": true
        },
        "gpu_ids": [
          0
        ],
        "num_epochs": 10,
        "num_gpus": 1,
        "num_nodes": 1,
        "optim": {
          "lr": 6e-05,
          "momentum": 0.9,
          "monitor_name": "val_loss",
          "optim": "adamw",
          "policy": "linear",
          "weight_decay": 0.01
        },
        "pretrained_model_path": "",
        "results_dir": "",
        "resume_training_checkpoint_path": "",
        "seed": 1234,
        "segment": {
          "loss": "ce",
          "weights": [
            0.5,
            0.5,
            0.5,
            0.8,
            1.0
          ]
        },
        "sync_batchnorm": false,
        "tensorboard": {
          "enabled": false,
          "infrequent_logging_frequency": 2
        },
        "use_distributed_sampler": false,
        "validation_interval": 1
      },
      "popular": [
        "validation_interval",
        "checkpoint_interval",
        "num_gpus",
        "num_nodes",
        "num_epochs",
        "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"
        },
        "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"
        },
        "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"
        },
        "optim": {
          "automl_default_parameters": [
            "train.optim.lr",
            "train.optim.momentum",
            "train.optim.weight_decay"
          ],
          "automl_enabled": false,
          "default": {
            "lr": 6e-05,
            "momentum": 0.9,
            "monitor_name": "val_loss",
            "optim": "adamw",
            "policy": "linear",
            "weight_decay": 0.01
          },
          "properties": {
            "lr": {
              "automl_enabled": true,
              "default": 6e-05,
              "description": "Optimizer learning rate",
              "maximum": Infinity,
              "minimum": 0.0,
              "type": "float"
            },
            "momentum": {
              "automl_enabled": true,
              "default": 0.9,
              "description": "The momentum (beta1) for the AdamW optimizer.",
              "maximum": 1.0,
              "minimum": 0.0,
              "title": "momentum - AdamW (beta1)",
              "type": "float"
            },
            "monitor_name": {
              "default": "val_loss",
              "description": "Monitor Name",
              "type": "string"
            },
            "optim": {
              "default": "adamw",
              "description": "Optimizer",
              "enum": [
                "adamw",
                "adam",
                "sgd"
              ],
              "type": "categorical"
            },
            "policy": {
              "default": "linear",
              "description": "Optimizer policy",
              "enum": [
                "linear",
                "step"
              ],
              "type": "categorical"
            },
            "weight_decay": {
              "automl_enabled": true,
              "default": 0.01,
              "description": "The weight decay coefficient.",
              "maximum": 1.0,
              "minimum": 0.0,
              "title": "weight decay",
              "type": "float"
            }
          },
          "type": "collection"
        },
        "pretrained_model_path": {
          "default": "",
          "description": "Pretrained model path",
          "title": "pretrained model path",
          "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"
        },
        "segment": {
          "automl_disabled_parameters": [
            "train.segment.weights"
          ],
          "automl_enabled": false,
          "default": {
            "loss": "ce",
            "weights": [
              0.5,
              0.5,
              0.5,
              0.8,
              1.0
            ]
          },
          "properties": {
            "loss": {
              "default": "ce",
              "description": "ChangeNet Segment loss",
              "enum": [
                "ce"
              ],
              "type": "categorical"
            },
            "weights": {
              "automl_enabled": false,
              "default": [
                0.5,
                0.5,
                0.5,
                0.8,
                1.0
              ],
              "description": "Multi-scale Segment loss weight",
              "type": "list"
            }
          },
          "type": "collection"
        },
        "sync_batchnorm": {
          "default": false,
          "description": "Enable synchronized batch normalization for multi-GPU training",
          "title": "sync_batchnorm",
          "type": "bool"
        },
        "tensorboard": {
          "automl_enabled": false,
          "default": {
            "enabled": false,
            "infrequent_logging_frequency": 2
          },
          "properties": {
            "enabled": {
              "default": false,
              "description": "Flag to enable tensorboard",
              "type": "bool"
            },
            "infrequent_logging_frequency": {
              "default": 2,
              "description": "infrequent_logging_frequency",
              "maximum": Infinity,
              "minimum": 0,
              "type": "int"
            }
          },
          "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": "evaluate",
    "core_module": "segformer",
    "model": "segformer",
    "network_arch": "segformer",
    "schema_action": "evaluate",
    "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, Needs Node.js

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-segformer · commit fce3cca

Licence

Apache License 2.0 (redistribution allowed)

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

Synced

2026-11-08

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-segformer.zip