Katsuya Hyodo

>python openvino_tiny-yolov3_test.py -d MYRIAD >--> get FPS=13 Correct result. >python openvino_tiny-yolov3_test.py -d CPU >--> get FPS=35 Correct result. OpenVINO is fully optimized for Intel CPUs with Intel TBB and Intel...

**https://github.com/PINTO0309/OpenVINO-YoloV3/tree/b5caa1c2117749b84719c0cd9750c10c452ef471/cpp#recompile** Please read my **`main.cpp`**. The Intel sample program and my program are different. **https://github.com/PINTO0309/OpenVINO-YoloV3/blob/b5caa1c2117749b84719c0cd9750c10c452ef471/cpp/main.cpp#L308-L310**

@vtpl1 **https://github.com/PINTO0309/OpenVINO-YoloV3/blob/master/script.txt** ```bash sudo python3 /opt/intel/computer_vision_sdk/deployment_tools/model_optimizer/mo_tf.py \ --input_model pbmodels/frozen_yolo_v3.pb \ --output_dir lrmodels/YoloV3/FP32 \ --data_type FP32 \ --batch 1 \ --tensorflow_use_custom_operations_config yolo_v3_changed.json sudo python3 /opt/intel/computer_vision_sdk/deployment_tools/model_optimizer/mo_tf.py \ --input_model pbmodels/frozen_yolo_v3.pb \ --output_dir lrmodels/YoloV3/FP16...

@vtpl1 I have not tried it with GPU yet. If you allow, other engineers are very pleased if you give feedback on your benchmark results to this issues.

@AllenDun >Have you checked the inference accuracy for the converted yolov3 model? Unfortunately, I have not confirmed it in detail yet. However, it feels pretty inferior in terms of experience....

@AllenDun, @deblauwetom Thank you for providing the information. I can not find a valid solution so far. :cry:

- Reference isssue https://github.com/mystic123/tensorflow-yolo-v3/issues/50 >Have you tried another way to convert Darknet? I have not tried it yet. I think I will search for it.

I began regenerating the model with reference to the following repository. So far, the loss value is steadily decreasing. **https://github.com/khanh1412/tiny-yolo-tensorflow.git** - ms-coco - pure Tensorflow model #2, #4

Hello everyone. I solved the problem of low precision. There was a mistake in the logic of preprocessing and postprocessing.

Thank you for proposing an improvement proposal. I will try when I have time.