search

bubbliiiing / mobilenet-ssd-keras

这是一个mobilenet-ssd-keras的源码，可以用于训练自己的轻量级ssd模型。
MIT License
105 stars 34 forks source link

mobilenet_ssd_weights 有问题 #2

Open wavelet2008 opened 4 years ago

wavelet2008 commented 4 years ago

"model_path": 'model_data/mobilenet_ssd_weights.h5'

net['predictions'] = concatenate([net['mbox_loc'], net['mbox_conf'], net['mbox_priorbox']], axis=2, name='predictions')

ValueError: A Concatenate layer requires inputs with matching shapes except for the concat axis. Got inputs shapes: [(None, 8732, 4), (None, 8732, 21), (None, 198144, 8)]

bubbliiiing commented 4 years ago

啥时候出的问题

peasfarmer commented 3 years ago

遇到一样的问题

peasfarmer commented 3 years ago

知道了，适配时ssd_layers.py改错了， if K.image_dim_ordering() == 'tf': 应该改成： if K.image_data_format() == 'channels_last': 如果写成了channels_first就会报这个错误

bubbliiiing commented 3 years ago

啊？我忘了我原来是啥了……是你修改后的错还是我原来的错

peasfarmer commented 3 years ago

不是你的错，修改支持tensorflow2时改错了

bubbliiiing commented 3 years ago

好的