search

ylshaooo / keras-zero-shot-detection

Keras implementation of zero-shot detection based on YOLOv3 model.
45 stars 15 forks source link

How to train using embedding file? #7

Closed jjwwczy closed 5 years ago

jjwwczy commented 5 years ago

I try to replace the 64-d attributes with the 300-d embedding, but it doesn't work,and get result as follows. tensorflow.python.framework.errors_impl.InvalidArgumentError: Input to reshape is a tensor with 10816 values, but the requested shape requires a multiple of 50700 [[Node: yolo_loss/Reshape_3 = Reshape[T=DT_FLOAT, Tshape=DT_INT32, _class=["loc:@training/Adam/gradients/yolo_loss/Reshape_3_grad/Reshape"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](yolo_loss/strided_slice_12, yolo_loss/Reshape_3/shape)]] [[Node: yolo_loss/while_2/LoopCond/_2615 = _HostRecvclient_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_7346_yolo_loss/while_2/LoopCond", tensor_type=DT_BOOL, _device="/job:localhost/replica:0/task:0/device:CPU:0"]] Where should I change?

wangshouxu commented 2 years ago

hi,I also try to use 300-d embedding, and has same question, where to change, thanks.