vis4d.zoo.base.models.yolox

YOLOX base model config.

Functions

get_model_setting(model_type)

Get YOLOX model setting.

get_yolox_callbacks_cfg(switch_epoch[, ...])

Get YOLOX callbacks for training.

get_yolox_cfg(num_classes, model_type[, ...])

Return default config for YOLOX model and loss.

get_yolox_model_cfg(num_classes, model_type)

Get YOLOX model.

get_yolox_optimizers_cfg(lr, num_epochs, ...)

Construct optimizer for YOLOX training.

get_model_setting(model_type)[source]

Get YOLOX model setting.

Return type:

tuple[float, float, int, list[int]]

get_yolox_callbacks_cfg(switch_epoch, shape=(480, 480), num_sizes=11, use_ema=True)[source]

Get YOLOX callbacks for training.

Return type:

list[ConfigDict]

get_yolox_cfg(num_classes, model_type, use_ema=True, weights=None)[source]

Return default config for YOLOX model and loss.

Parameters:
  • num_classes (FieldReference | int) – Number of classes.

  • model_type (str) – Model type. Must be one of ‘tiny’, ‘small’, ‘large’, ‘xlarge’.

  • use_ema (bool, optional) – Whether to use EMA. Defaults to True.

  • weights (str | None, optional) – Weights to load. Defaults to None.

Return type:

tuple[ConfigDict, ConfigDict]

get_yolox_model_cfg(num_classes, model_type)[source]

Get YOLOX model.

Return type:

ConfigDict

get_yolox_optimizers_cfg(lr, num_epochs, warmup_epochs, num_last_epochs)[source]

Construct optimizer for YOLOX training.

Return type:

list[OptimizerConfig]