JunweiLiang / Object_Detection_Tracking

Out-of-the-box code and models for CMU's object detection and tracking system for multi-camera surveillance videos. Speed optimized Faster-RCNN model. Tensorflow based. Also supports EfficientDet. WACVW'20
MIT License
479 stars 135 forks source link

Issues about NCHW and in_channel #14

Closed Neo-YH closed 4 years ago

Neo-YH commented 4 years ago

image 学长,我执行代码时报了default maxpooling only support NHWC的问题,我看网上的解决方案把NCHW全替换成了NHWC。这个报错消失了,但是接下来Assert in_channel is not None出问题了。 image 我看我把NCHW全部替换之后channel_axis就只能等于3了,in_shape取出来的就是None。不太理解这个过程,我应该怎么修复这个报错啊。万分感谢学长。

JunweiLiang commented 4 years ago

第一个错误看起来是你的TF没有装GPU版本,你看一下下面代码的输出是否0: print("Num GPUs Available: ", len(tf.config.experimental.list_physical_devices('GPU'))) 改成NHWC的话,很多代码都需要相应改变进行transpose操作,不建议改。

Neo-YH commented 4 years ago

感谢学长,确实是GPU的问题,重装了CUDA,cuDNN对应的tensorflow-gpu之后报错就解决了 image