torchsig.models.model_utils.model_utils_1d.conversions_to_1d

Functions

avgpool2d_to_avgpool1d	returns a MaxPool1d layer corresponding to the input MaxPool2d layer no muation is performed
batchNorm2d_to_GBN1d	returns a GBN1d [Ghost Batch Norm] layer corresponding to the input BatchNorm2d layer no muation is performed
batchNorm2d_to_batchNorm1d	returns a BatchNorm1d layer corresponding to the input BatchNorm2d layer no muation is performed
conv2d_to_conv1d	returns a 1d conv layer corresponding to the input conv2d layer no muation is performed
convert_2d_model_to_1d	converts a 2d model to a corresponding 1d model by replacing convolutional layers and other 2d layers with their 1d equivalents experimental; may not fully convert models with unrecognized layer types mutates input model; returns the mutated model this function is still under development and may not correctly convert all 2d layer types, or may have unexpected behavior on models that perform reshaping internally
get_1d_kernel	adjusts kernel shapes from 2d to 1d; returns 1d shape (single int)
make_fast_avg_pooling_layer	returns a FastGlobalAvgPool1d layer
maxpool2d_to_maxpool1d	returns a MaxPool1d layer corresponding to the input MaxPool2d layer no muation is performed
squeezeExcite_to_squeezeExcite1d	returns a GBN1d [Ghost Batch Norm] layer corresponding to the input BatchNorm2d layer no muation is performed
try_default	tries to return the result of a function call; returns result_on_error on error