tensorflow-yolov4

NOTICE: This is a fork from https://github.com/hunglc007/tensorflow-yolov4-tflite, to add support to convert YOLO .weights to tf's .pb or tf.keras's .h5 format.

YOLOv4 Implemented in Tensorflow 2.0. Convert YOLO v4 .weights to .pb and .tflite format for tensorflow and tensorflow lite.

YOLOv4, YOLOv4-tiny Implemented in Tensorflow 2.0. Convert YOLO v4, YOLOv3, YOLO tiny .weights to .pb, .tflite and trt format for tensorflow, tensorflow lite, tensorRT.

Download yolov4.weights file: https://drive.google.com/open?id=1cewMfusmPjYWbrnuJRuKhPMwRe_b9PaT

Prerequisites

• Tensorflow 2.3.0rc0

Performance

Demo

# Convert darknet weights to tensorflow
## yolov4
python save_model.py --weights ./data/yolov4.weights --output ./checkpoints/yolov4-416 --input_size 416 --model yolov4 

## yolov4-tiny
python save_model.py --weights ./data/yolov4-tiny.weights --output ./checkpoints/yolov4-tiny-416 --input_size 416 --model yolov4 --tiny

# Run demo tensorflow
python detect.py --weights ./checkpoints/yolov4-416 --size 416 --model yolov4 --image ./data/kite.jpg

python detect.py --weights ./checkpoints/yolov4-tiny-416 --size 416 --model yolov4 --image ./data/kite.jpg --tiny

If you want to run yolov3 or yolov3-tiny change --model yolov3 in command

Output

Yolov4 original weight

Yolov4 tflite int8

Convert to .h5

# yolov4
python convert.py --weights ./data/yolov4.weights --output ./data/yolov4.h5

Convert to .pb

# yolov4
python convert.py --weights ./data/yolov4.weights --output ./data/yolov4-pb

Convert to .tflite

# Save tf model for tflite converting
python save_model.py --weights ./data/yolov4.weights --output ./checkpoints/yolov4-416 --input_size 416 --model yolov4 --framework tflite

# yolov4
python convert_tflite.py --weights ./checkpoints/yolov4-416 --output ./checkpoints/yolov4-416.tflite

# yolov4 quantize float16
python convert_tflite.py --weights ./checkpoints/yolov4-416 --output ./checkpoints/yolov4-416-fp16.tflite --quantize_mode float16

# yolov4 quantize int8
python convert_tflite.py --weights ./checkpoints/yolov4-416 --output ./checkpoints/yolov4-416-int8.tflite --quantize_mode int8 --dataset ./coco_dataset/coco/val207.txt

# Run demo tflite model
python detect.py --weights ./checkpoints/yolov4-416.tflite --size 416 --model yolov4 --image ./data/kite.jpg --framework tflite

Yolov4 and Yolov4-tiny int8 quantization have some issues. I will try to fix that. You can try Yolov3 and Yolov3-tiny int8 quantization

Convert to TensorRT

python save_model.py --weights ./data/yolov3.weights --output ./checkpoints/yolov3.tf --input_size 416 --model yolov3
python convert_trt.py --weights ./checkpoints/yolov3.tf --quantize_mode float16 --output ./checkpoints/yolov3-trt-fp16-416

# yolov3-tiny
python save_model.py --weights ./data/yolov3-tiny.weights --output ./checkpoints/yolov3-tiny.tf --input_size 416 --tiny
python convert_trt.py --weights ./checkpoints/yolov3-tiny.tf --quantize_mode float16 --output ./checkpoints/yolov3-tiny-trt-fp16-416

# yolov4
python save_model.py --weights ./data/yolov4.weights --output ./checkpoints/yolov4.tf --input_size 416 --model yolov4
python convert_trt.py --weights ./checkpoints/yolov4.tf --quantize_mode float16 --output ./checkpoints/yolov4-trt-fp16-416

Evaluate on COCO 2017 Dataset

# run script in /script/get_coco_dataset_2017.sh to download COCO 2017 Dataset
# preprocess coco dataset
cd data
mkdir dataset
cd ..
cd scripts
python coco_convert.py --input ./coco/annotations/instances_val2017.json --output val2017.pkl
python coco_annotation.py --coco_path ./coco 
cd ..

# evaluate yolov4 model
python evaluate.py --weights ./data/yolov4.weights
cd mAP/extra
python remove_space.py
cd ..
python main.py --output results_yolov4_tf

mAP50 on COCO 2017 Dataset

Detection	512x512	416x416	320x320
YoloV3	55.43	52.32
YoloV4	61.96	57.33

Benchmark

python benchmarks.py --size 416 --model yolov4 --weights ./data/yolov4.weights

TensorRT performance

YoloV4 416 images/s	FP32	FP16	INT8
Batch size 1	55	116
Batch size 8	70	152

Tesla P100

Detection	512x512	416x416	320x320
YoloV3 FPS	40.6	49.4	61.3
YoloV4 FPS	33.4	41.7	50.0

Tesla K80

Detection	512x512	416x416	320x320
YoloV3 FPS	10.8	12.9	17.6
YoloV4 FPS	9.6	11.7	16.0

Tesla T4

Detection	512x512	416x416	320x320
YoloV3 FPS	27.6	32.3	45.1
YoloV4 FPS	24.0	30.3	40.1

Tesla P4

Detection	512x512	416x416	320x320
YoloV3 FPS	20.2	24.2	31.2
YoloV4 FPS	16.2	20.2	26.5

Macbook Pro 15 (2.3GHz i7)

Detection	512x512	416x416	320x320
YoloV3 FPS
YoloV4 FPS

Traning your own model

# Prepare your dataset
# If you want to train from scratch:
In config.py set FISRT_STAGE_EPOCHS=0 
# Run script:
python train.py

# Transfer learning: 
python train.py --weights ./data/yolov4.weights

The training performance is not fully reproduced yet, so I recommended to use Alex's Darknet to train your own data, then convert the .weights to tensorflow or tflite.

TODO

• Convert YOLOv4 to TensorRT
• YOLOv4 tflite on android
• YOLOv4 tflite on ios
• Training code
• Update scale xy
• ciou
• Mosaic data augmentation
• Mish activation
• yolov4 tflite version
• yolov4 in8 tflite version for mobile

References

• YOLOv4: Optimal Speed and Accuracy of Object Detection YOLOv4.
• darknet

My project is inspired by these previous fantastic YOLOv3 implementations:

• Yolov3 tensorflow
• Yolov3 tf2