Open Ignotus opened 5 years ago
Hi, The LCNN29 network implementation is not exactly same as the paper. For MFM_fc1, we have implemented it and used it for training, but the result was slight worse that not using. You can try to compare the different by adding MFM_fc layer after mod.fcnn_layer(512) layer.
Hello,
Thank you for the effort to port it to Tensorflow.
In the original paper fc1 is followed by MFM_fc1. I see you implemented MFMfc but it's not used anywhere. Is it a bug or intentionally made decision?