Universite Paris-Est Marne-la-Vallee IMAGINE/LIGM torch neural network routines
Following modules are here for now:
inn.SpatialMaxPooling(kW,kH,dW,dH)
inn.SpatialAveragePooling(kW,kH,dW,dH)
inn.SpatialCrossResponseNormalization(size, [alpha = 0.0001], [beta = 0.75], [k = 1])
inn.LocalResponseNormalization([size = 3], [alpha = 0.00005], [beta = 0.75])
inn.MeanSubtraction(mean)
inn.SpatialPyramidPooling({{w1,h1},{w2,h2},...,{wn,hn}})The difference with inn.SpatialMax(Average)Pooling and nn.SpatialMax(Average)Pooling is that output size computed with ceil instead of floor (as in Caffe and cuda-convnet2). Also SpatialAveragePooling does true average pooling, meaning that it divides outputs by kW*kH.
inn.SpatialMax(Average)Pooling(kW,kH,dW,dH) is equal to cudnn.SpatialMax(Average)Pooling(kW,kH,dW,dH):ceil().
inn.SpatialCrossResponseNormalization is local response normalization across maps in BDHW format (thanks to Caffe!). For details refer to https://code.google.com/p/cuda-convnet/wiki/LayerParams#Local_response_normalization_layer_(across_maps)
inn.LocalResponseNormalization is a local response normalization in the same map in BDHW format. For details refer to https://code.google.com/p/cuda-convnet/wiki/LayerParams#Local_response_normalization_layer_(same_map)
inn.MeanSubtraction(mean) is done to subtract the Imagenet mean directly on GPU. Mean tensor is expanded to BDHW batches without using additional memory.
inn.SpatialPyramidPooling({{w1,h1},{w2,h2},...,{wn,hn}}) is a pyramid of regions obtained by using Spatial Adaptive Max Pooling with parameters (w1,h1),...,(wn,hn) in the input. The result is a fixed-sized vector of size w1*h1*...wn*hn for any input dimension. For details see http://arxiv.org/abs/1406.4729