ssd_resnet50_v1_fpn_shared_box_predictor_640x640_coco14_sync_2018_07_03模型下載地址：TensorFlow 1 Detection Model Zoo

OpenVINO版本：2021.3.394

第一步：新建tf1_ov虛擬環(huán)境，注意Python=3.6。選擇3.7或3.8會導(dǎo)致install_prerequisites_tf.bat安裝tensorflow版本出錯。在pip庫中，tensorflow 1.15支持的Python版本最高為3.6

第二步：運行install_prerequisites_tf.bat，配置model optimizer的運行環(huán)境

第三步，執(zhí)行命令：python mo_tf.py --input_model d:\ov_mo\ssd_resnet50_v1_fpn_shared_box_predictor_640x640_coco14_sync_2018_07_03\frozen_inference_graph.pb --tensorflow_object_detection_api_pipeline_config d:\ov_mo\ssd_resnet50_v1_fpn_shared_box_predictor_640x640_coco14_sync_2018_07_03\pipeline.config --output_dir d:\ov_mo\ssd_resnet50_v1_fpn_shared_box_predictor_640x640_coco14_sync_2018_07_03\ --model_name ssd_resnet50_tf1_ov --data_type FP16 --reverse_input_channels --transformations_config extensions\front\tf\ssd_v2_support.json

ssd_resnet50_v1_fpn.png

參考文獻：model_optimizer_args