BiteSizedChunks.comLearn one small thing at a time.

Course contentsShow

Machine Learning and Deep Learning

851What is Convolution: Mathematical Definition
852Convolution as a Sliding Window
853Kernels and Filters: Terminology
8542D Convolution for Images
855Stride: Controlling Step Size
856Padding: Zero, Valid, and Same
857Computing Output Dimensions
858Multi-Channel Convolution
859Multiple Output Channels
860Parameter Count in Convolutional Layers
861Local Connectivity and Weight Sharing
862Translation Equivariance
863Common Filter Sizes: 3x3, 5x5, 1x1
864Dilated Convolution
865Grouped Convolution
866Depthwise Separable Convolution
867Why Pooling? Spatial Downsampling and Invariance
868Max Pooling Operation
869Average Pooling Operation
870Pooling Hyperparameters: Kernel Size and Stride
871Pooling vs Strided Convolutions
872Global Average Pooling
873Batch Normalization in CNNs
874Dropout for CNNs: Spatial Dropout
8751x1 Convolutions: Bottleneck Layers
876Activation Functions in CNN Architectures
877Building Blocks: Conv-BN-ReLU Patterns
878Fully Connected Layers as Classification Heads
879What is a Receptive Field?
880Calculating Receptive Fields in Sequential Layers
881Receptive Field Formula
882Impact of Stride on Receptive Fields
883Pooling Layers and Receptive Field Expansion
884Dilated Convolutions for Large Receptive Fields
885Effective vs Theoretical Receptive Fields
886Network Depth and Feature Hierarchy
887Receptive Fields in Modern Architectures
888Designing Networks with Receptive Field Constraints

Machine Learning and Deep Learning

851What is Convolution: Mathematical Definition
852Convolution as a Sliding Window
853Kernels and Filters: Terminology
8542D Convolution for Images
855Stride: Controlling Step Size
856Padding: Zero, Valid, and Same
857Computing Output Dimensions
858Multi-Channel Convolution
859Multiple Output Channels
860Parameter Count in Convolutional Layers
861Local Connectivity and Weight Sharing
862Translation Equivariance
863Common Filter Sizes: 3x3, 5x5, 1x1
864Dilated Convolution
865Grouped Convolution
866Depthwise Separable Convolution
867Why Pooling? Spatial Downsampling and Invariance
868Max Pooling Operation
869Average Pooling Operation
870Pooling Hyperparameters: Kernel Size and Stride
871Pooling vs Strided Convolutions
872Global Average Pooling
873Batch Normalization in CNNs
874Dropout for CNNs: Spatial Dropout
8751x1 Convolutions: Bottleneck Layers
876Activation Functions in CNN Architectures
877Building Blocks: Conv-BN-ReLU Patterns
878Fully Connected Layers as Classification Heads
879What is a Receptive Field?
880Calculating Receptive Fields in Sequential Layers
881Receptive Field Formula
882Impact of Stride on Receptive Fields
883Pooling Layers and Receptive Field Expansion
884Dilated Convolutions for Large Receptive Fields
885Effective vs Theoretical Receptive Fields
886Network Depth and Feature Hierarchy
887Receptive Fields in Modern Architectures
888Designing Networks with Receptive Field Constraints

← Machine Learning and Deep Learning

Lesson 875 of 3,538·20. Convolutional Neural Networks: FundamentalsPro lesson

1x1 Convolutions: Bottleneck Layers

Learn how 1x1 convolutions change channel dimensionality without affecting spatial size, enabling efficient bottleneck architectures.

This lesson is for subscribers

You've completed the free preview. Subscribe to unlock every lesson in every course.

See pricing Back to course