Skip to content

Latest commit

 

History

History
41 lines (39 loc) · 1.79 KB

README.md

File metadata and controls

41 lines (39 loc) · 1.79 KB

Introduction

Translating the model in yolox to tensorflow2.0.
you should download yolox model: yolox_s, yolox_m, yolox_l, yolox_x, yolox_nano, yolox_tiny in exp folder( see the content ).
Supported tensorflow: tensorflow saved_model, tflite

Content

  • yolox_torch2tf
    • datasets
      • datasets.py
    • exp
      • export.py
      • para.py
    • models
      • tf_darknet.py
      • tf_network_blocks.py
      • tf_yolo_head.py
      • tf_yolo_pafpn.py

yolox_torch2tf: root.
datasets: to export tflite int8, download coco128 there as .\datasets\coco128.
models: model translation code with tensorflow.
exp: export code and save tf model, you also need download yolox model to this folder.

Usage

parameter int8 only for tflite int8, otherwise please ignore it.

    python path/to/export.py -n yolox_nano --tsize 640 --include saved_model --device cpu --int8
                                yolox_tiny                       tflite               gpu
                                yolox_s
                                yolox_x
                                yolox_m
                                yolox_l