yolo_quantization

yolo_quantization copied to clipboard

batch = 1 subdivisions=1才能运行

5

# Training batch=1 subdivisions=1 其他的都会报CUDA错误

mobilejammer

darknet: ./src/blas.c:66: fake_quant_with_min_max_channel: Assertion `0' failed. 已放弃 (核心已转储)

1

如题

mobilejammer

不知道程序是怎么读入数据的，在voc_nok.data有定义训练的txt，但是label文件，image文件怎么组织呢

mobilejammer

这个实现其他网络来进行训练量化嘛？还是只针对于yolov3-tiny这个网络

shadows-z

Do you perform quantitative operations during train according to the article?

1

hello, I read your code and can't find the quantitative operations during train ,can your tell me which function do this operations? Thank you!

yanxiaopan

你好，请问有加入YOLOv5的计划吗？

1

YOLOv5速度更高，精度更好，期待v5的实现！

Wulingtian