deepdetect-pytorch-worker

Port external PyTorch object detection models into DeepDetect by generating adapters, mapping checkpoints, and converting label and box formats to match DeepDetect detection schemas using the external-pytorch-detector profile.

Yes, existing PyTorch checkpoints work with the DeepDetect external worker backend by applying checkpoint compatibility validation and mapping model factories and configs into DeepDetect-compatible workflows.

DeepDetect handles label and box conversion by applying explicit translation rules between upstream PyTorch model outputs and DeepDetect detection schemas, ensuring prediction format alignment during inference.

Validate training and inference for ported PyTorch detection models by testing dataset loading, connector pull flows, prediction conversion, and checkpoint compatibility through repeatable CLI-based workflows.

You must handle explicit dependency configuration when connecting a custom PyTorch detector to DeepDetect, ensuring generic worker contracts are met for correct model execution and inference workflows.

PyTorch detector predictions misalign with the DeepDetect detection schema when adapter generation lacks correct translation between upstream model outputs and DeepDetect label and box format conventions.