FiK^ddlmZddlmZddlZddlmZddlmZddl m Z Gdd e Z y) ) annotations)PathN)LOGGER)check_executorch_requirements) BaseBackendc eZdZdZddZddZy)ExecuTorchBackendzMeta ExecuTorch inference backend for on-device deployment. Loads and runs inference with Meta ExecuTorch models (.pte files) using the ExecuTorch runtime. Supports both standalone .pte files and directory-based model packages with metadata. ctjd|dtddlm}t |}|j r t|jd}|dz }n|}|jdz }|jjt|}|jd|_|jr'ddlm}|j%|j'|y y ) zLoad an ExecuTorch model from a .pte file or directory. Args: weight (str | Path): Path to the .pte model file or directory containing the model. zLoading z for ExecuTorch inference...r)Runtimez*.ptez metadata.yamlforward)YAMLN)rinforexecutorch.runtimer ris_dirnextrglobparentget load_programstr load_methodmodelexistsultralytics.utilsrapply_metadataload)selfweightr w model_file metadata_fileprogramrs d/var/www/pod-logistic/pod-ai/venv/lib/python3.12/site-packages/ultralytics/nn/backends/executorch.py load_modelzExecuTorchBackend.load_models  hvh&BCD%'. L 88:aggg./J/MJHH6M++-,,S_=((3     ! .    - 8 9 "c:|jj|gS)zRun inference using the ExecuTorch runtime. Args: im (torch.Tensor): Input image tensor in BCHW format, normalized to [0, 1]. Returns: (list): Model predictions as a list of ExecuTorch output values. )rexecute)rims r$r zExecuTorchBackend.forward2szz!!2$''r&N)rz str | PathreturnNone)r)z torch.Tensorr*list)__name__ __module__ __qualname____doc__r%r r&r$r r s :8 (r&r ) __future__rpathlibrtorchrrultralytics.utils.checksrbaserr r1r&r$r7s%# $B,( ,(r&