# losses

Python API: `luxonis_train.attached_modules.losses`

Loss functions, grouped by the task they supervise.

 * classification and segmentation:
   [CrossEntropyLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/cross_entropy.md),
   [BCEWithLogitsLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/bce_with_logits.md),
   [SmoothBCEWithLogitsLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/smooth_bce_with_logits.md),
   [SigmoidFocalLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/sigmoid_focal_loss.md),
   [SoftmaxFocalLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/softmax_focal_loss.md),
   and
   [OHEMLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/ohem_loss.md)
 * bounding boxes:
   [AdaptiveDetectionLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/adaptive_detection_loss.md)
   and
   [PrecisionDFLDetectionLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/precision_dfl_detection_loss.md)
 * instance keypoints:
   [EfficientKeypointBBoxLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/efficient_keypoint_bbox_loss.md)
 * FOMO object centers:
   [FOMOLocalizationLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/fomo_localization_loss.md)
 * instance segmentation:
   [PrecisionDFLSegmentationLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/precision_dfl_segmentation_loss.md)
 * anomaly detection:
   [ReconstructionSegmentationLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/reconstruction_segmentation_loss.md)
 * OCR:
   [CTCLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/ctc_loss.md)
 * embeddings: the [pytorch-metric-learning](https://kevinmusgrave.github.io/pytorch-metric-learning/losses/) losses, wrapped by
   [EmbeddingLossWrapper](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/embedding_losses.md)

A config attaches a loss to a node in the `losses` list of the node. The `Compatible with` section of each loss lists the nodes
that it accepts.
[BaseLoss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/base_loss.md)
describes how to write a new loss. The training step raises `ValueError` when no node has a loss.

## Child Pages

 * [adaptive_detection_loss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/adaptive_detection_loss.md)
 * [base_loss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/base_loss.md)
 * [bce_with_logits](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/bce_with_logits.md)
 * [cross_entropy](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/cross_entropy.md)
 * [ctc_loss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/ctc_loss.md)
 * [efficient_keypoint_bbox_loss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/efficient_keypoint_bbox_loss.md)
 * [embedding_losses](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/embedding_losses.md)
 * [fomo_localization_loss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/fomo_localization_loss.md)
 * [ohem_loss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/ohem_loss.md)
 * [precision_dfl_detection_loss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/precision_dfl_detection_loss.md)
 * [precision_dfl_segmentation_loss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/precision_dfl_segmentation_loss.md)
 * [reconstruction_segmentation_loss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/reconstruction_segmentation_loss.md)
 * [sigmoid_focal_loss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/sigmoid_focal_loss.md)
 * [smooth_bce_with_logits](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/smooth_bce_with_logits.md)
 * [softmax_focal_loss](https://docs.luxonis.com/software-v3/ai-inference/model-source/training/luxonis-train/luxonis-train-api-reference/attached_modules/losses/softmax_focal_loss.md)