Hi,
I want to train one model which includes VGG net and LSTM module. The LSTM module is connected to VGG net to do activity recognition. Can DIGITS be used to train it?
As we all known, the special point is the clip marker input. And I have no idea about the input format of clip marker to create HDF5 data format. Some open-source codes about LSTM using python layer to create input source and they are a little complicated to understand. Can you give me some advice?
Thanks
Hi, I want to train one model which includes VGG net and LSTM module. The LSTM module is connected to VGG net to do activity recognition. Can DIGITS be used to train it? As we all known, the special point is the clip marker input. And I have no idea about the input format of clip marker to create HDF5 data format. Some open-source codes about LSTM using python layer to create input source and they are a little complicated to understand. Can you give me some advice? Thanks