DeepDetect v0.26.0

Features

• generate mask for diffusion models in trt (5c16ad0)
• opencv optional build (ce9b9a7)
• output: add false positive metrics for detection (bec49c4)
• torch: add JIT FusionStrategy selection (d2331be)
• torch: Added script to trace HuggingFace Transformers CLIP models based on https://huggingface.co/openai/clip-vit-large-patch14-336 (c2373ea)

Bug Fixes

• torch: errors in input connector are caught correctly (ac09c52)
• torch: segmentation with test_batch_size > 1 (0d8d3da)

Docker images:

• CPU version: docker pull docker.jolibrain.com/deepdetect_cpu:v0.26.0
• GPU (CUDA only): docker pull docker.jolibrain.com/deepdetect_gpu:v0.26.0
• GPU (CUDA and Tensorrt) :docker pull docker.jolibrain.com/deepdetect_cpu_tensorrt:v0.26.0
• GPU with torch backend: docker pull docker.jolibrain.com/deepdetect_gpu_torch:v0.26.0
• All images available from https://docker.jolibrain.com/, list images with {"repositories":["cuda12.5.1-devel-ubuntu22.04-preinst-devel","deepdetect_cpu","deepdetect_cpu_torch","deepdetect_gpu","deepdetect_gpu_tensorrt","deepdetect_gpu_torch","filebrowser","gpustat_server","joligen_server","joligen_ui","jupyter_dd_notebook","platform_annotations_backend","platform_annotations_frontend","platform_data","platform_ui"]}