How to run inference yolov5.tflite that quantized int8?

[Mahmoud Alipour]

Hi everyone, I'm going to obtain bounding box of tflite of yolov5 that is int8 quantized. How can I find that?

I get bounding boxes of yolov5 with fp12 quantization, with the code in this link. but it does not work on int8 quantization.

  

[Jiacong Fang]

You could run the inference with int8 tflite models with

```

python detect.py --weights yolov5s-int8.tflite

```

using the master branch of yolov5.

More information can be found at https://github.com/ultralytics/yolov5/pull/1127