mobilenetv3-yolov3

mobilenetv3-yolov3 copied to clipboard

预训练权重我使用的是mobilenetv3分类集上的权重，但是map却很低，是冻结层数不对所导致的吗

1

xingstart

How to account the mAP of the trained model?

2

Thanks for your great work, could you tell me how to get the AP of trained model?

ShihuaiXu

could you tell me the final result in voc?I used the mobilenetv3 pretrained model,but the result is still not good

1

chengyequn

can you provide the pretrain model again?The previous URL cannot be opened

1

JiliuLiSU

No module named 'yolo3.mobilenet_base'

1

ModuleNotFoundError: No module named 'yolo3.mobilenet_base'

HMAJG

Can you provide the model you have trained?

3

Hi, can you provide the model you have trained? Thanks!

caoabc

model on tiny

4

du you have pretrianed model on tiny ? @tanluren

Jenny0420

这个问题怎么解决：ValueError: Layer #220 (named "conv2d_35"), weight <tf.Variable 'conv2d_35/kernel:0' shape=(1, 1, 160, 18) dtype=float32> has shape (1, 1, 160, 18), but the saved weight has shape (75, 160, 1, 1).

3

lucyLLLL

Could you tell me the version of tensorflow and keras, thank you

1

maxwellTURE

convert to keras , benchmark model

2

Thanks for your model. I have some questions to ask you. - You can tell me clearly how to convert the h5 keras to pb tensorflow model and how to...

phucnhs