LeNet

ConvNet

class lucid.models.LeNet(config: LeNetConfig)

Overview

The LeNet base class provides a flexible implementation for defining various versions of the LeNet architecture, including LeNet-1, LeNet-4, and LeNet-5.

It takes a LeNetConfig object that centralizes the convolution, classifier, and activation settings for a concrete LeNet variant or a custom architecture.

        %%{init: {"flowchart":{"curve":"monotoneX","nodeSpacing":50,"rankSpacing":50}} }%%
flowchart LR
  linkStyle default stroke-width:2.0px
  subgraph sg_m0["<span style='font-size:20px;font-weight:700'>lenet_5</span>"]
  style sg_m0 fill:#000000,fill-opacity:0.05,stroke:#000000,stroke-opacity:0.75,stroke-width:1px
    subgraph sg_m1["conv1 x 2"]
      direction TB;
    style sg_m1 fill:#000000,fill-opacity:0.05,stroke:#000000,stroke-opacity:0.75,stroke-width:1px
      m1_in(["Input"]);
      m1_out(["Output"]);
  style m1_in fill:#e2e8f0,stroke:#64748b,stroke-width:1px;
  style m1_out fill:#e2e8f0,stroke:#64748b,stroke-width:1px;
      m2["Conv2d<br/><span style='font-size:11px;color:#c53030;font-weight:400'>(1,1,32,32) → (1,6,28,28)</span>"];
      m3["Tanh"];
      m4["AvgPool2d<br/><span style='font-size:11px;color:#b7791f;font-weight:400'>(1,6,28,28) → (1,6,14,14)</span>"];
    end
    m5["Linear<br/><span style='font-size:11px;color:#2b6cb0;font-weight:400'>(1,400) → (1,120)</span>"];
    m6["Tanh"];
    m7["Linear<br/><span style='font-size:11px;color:#2b6cb0;font-weight:400'>(1,120) → (1,84)</span>"];
    m8["Tanh"];
    m9["Linear<br/><span style='font-size:11px;color:#2b6cb0;font-weight:400'>(1,84) → (1,10)</span>"];
    m10["Tanh"];
  end
  input["Input<br/><span style='font-size:11px;color:#a67c00;font-weight:400'>(1,1,32,32)</span>"];
  output["Output<br/><span style='font-size:11px;color:#a67c00;font-weight:400'>(1,10)</span>"];
  style input fill:#fff3cd,stroke:#a67c00,stroke-width:1px;
  style output fill:#fff3cd,stroke:#a67c00,stroke-width:1px;
  style m2 fill:#ffe8e8,stroke:#c53030,stroke-width:1px;
  style m3 fill:#faf5ff,stroke:#6b46c1,stroke-width:1px;
  style m4 fill:#fefcbf,stroke:#b7791f,stroke-width:1px;
  style m5 fill:#ebf8ff,stroke:#2b6cb0,stroke-width:1px;
  style m6 fill:#faf5ff,stroke:#6b46c1,stroke-width:1px;
  style m7 fill:#ebf8ff,stroke:#2b6cb0,stroke-width:1px;
  style m8 fill:#faf5ff,stroke:#6b46c1,stroke-width:1px;
  style m9 fill:#ebf8ff,stroke:#2b6cb0,stroke-width:1px;
  style m10 fill:#faf5ff,stroke:#6b46c1,stroke-width:1px;
  input -.-> m2;
  m10 --> output;
  m1_in -.-> m2;
  m1_out --> m5;
  m2 --> m3;
  m3 --> m4;
  m4 --> m1_in;
  m4 --> m1_out;
  m5 --> m6;
  m6 --> m7;
  m7 --> m8;
  m8 --> m9;
  m9 --> m10;
    

Class Signature

class LeNet(nn.Module):
    def __init__(self, config: LeNetConfig) -> None

Parameters

• config (LeNetConfig) A configuration object describing the two convolution stages, the classifier layer sizes, the flattened input size for the first linear layer, and the activation module class used between layers.

Attributes

• config (LeNetConfig) The configuration used to build the model.

• conv1, conv2 (nn.Sequential) Sequential blocks containing convolution, activation, and average pooling.

• fc1, fc2, … Dynamically added fully connected classifier layers.

Methods

• forward(x: Tensor) -> Tensor Performs the forward pass through the convolution blocks, flattens the features, and applies the classifier stack.

  def forward(self, x):
      x = self.conv1(x)
      x = self.conv2(x)
      x = x.reshape(x.shape[0], -1)
      ...
      return x
  

Example Usage

Below is an example of defining and using a custom LeNet architecture with LeNetConfig:

import lucid.models as models
import lucid.nn as nn

config = models.LeNetConfig(
    conv_layers=[
        {"out_channels": 6},
        {"out_channels": 16},
    ],
    clf_layers=[120, 84, 10],
    clf_in_features=16 * 5 * 5,
    base_activation=nn.Tanh,
)
custom_lenet = models.LeNet(config)

# Sample input tensor (e.g., 32x32 grayscale image)
input_tensor = lucid.Tensor([...])

# Forward pass
output = custom_lenet(input_tensor)
print(output)