the merge problem in SSD document
Opened this issue · 3 comments
When I try to run the SSD.ipny , I got the error. "concat" mode can only merge layers with matching output shapes except for the concat axis. Layer shapes: [(None, 38, 38, 512), (None, 19, 19, 1024), (None, 10, 10, 512), (None, 5, 5, 256), (None, 3, 3, 256), (None, 1, 1, 256)]
The error point to line 274 ,where the code wants to concatenate the six layers's prior boxes.
Could you tell me the reason of error? thank you for much.
oh, I get the reason. the reason is the different version of Keras. The great work in this project is using the Keras version 1.22, where writing your own Keras layer design the layer's output shape by get_output_shape_for(input_shape). But there are some different in Keras version 2.1.3, where writing your own Keras layer design the layer's output shape by compute_output_shape(input_shape).
So, you just replace the old name of function by the new.
Hey ,
Actually i m facing the same issue and i am not able to resolve can you please help me with this.
Thank you
down keras version to 2.1.6