In order to perform inference on the trained model the input and output nodes for encode/decode must be known. Does anyone know how to figure this out / what the naming scheme is for a generic model with given hyperparameters? I've tried visualizing my trained model in tensorflow but the graph is extremely cluttered compared to that of the pretrained model making it impossible to decipher outputs / input nodes. Any help would be appreciated.
In order to perform inference on the trained model the input and output nodes for encode/decode must be known. Does anyone know how to figure this out / what the naming scheme is for a generic model with given hyperparameters? I've tried visualizing my trained model in tensorflow but the graph is extremely cluttered compared to that of the pretrained model making it impossible to decipher outputs / input nodes. Any help would be appreciated.