Source code for hyrax.models.hyrax_cnn
# ruff: noqa: D101, D102
# This example model is taken from the PyTorch CIFAR10 tutorial:
# https://pytorch.org/tutorials/beginner/blitz/cifar10_tutorial.html#define-a-convolutional-neural-network
import logging
import torch
import torch.nn as nn
import torch.nn.functional as F # noqa N812
from .model_registry import hyrax_model
[docs]
logger = logging.getLogger(__name__)
@hyrax_model
[docs]
class HyraxCNN(nn.Module):
"""
This CNN is designed to work with datasets that are prepared with Hyrax's HSC Data Set class.
"""
def __init__(self, config, data_sample=None):
super().__init__()
if data_sample is None:
raise ValueError("A `data_sample` must be provided to HyraxCNN for dynamic sizing.")
image_sample = data_sample[0]
self.num_input_channels, self.image_width, self.image_height = image_sample.shape
hidden_channels_1 = 6
hidden_channels_2 = 16
# Calculate how much our convolutional layers and pooling will affect
# the size of final convolution.
#
# If the number of layers are changed this will need to be rewritten.
conv1_end_w = self.conv2d_output_size(self.image_width, kernel_size=5)
conv1_end_h = self.conv2d_output_size(self.image_height, kernel_size=5)
pool1_end_w = self.pool2d_output_size(conv1_end_w, kernel_size=2, stride=2)
pool1_end_h = self.pool2d_output_size(conv1_end_h, kernel_size=2, stride=2)
conv2_end_w = self.conv2d_output_size(pool1_end_w, kernel_size=5)
conv2_end_h = self.conv2d_output_size(pool1_end_h, kernel_size=5)
pool2_end_w = self.pool2d_output_size(conv2_end_w, kernel_size=2, stride=2)
pool2_end_h = self.pool2d_output_size(conv2_end_h, kernel_size=2, stride=2)
[docs]
self.conv1 = nn.Conv2d(self.num_input_channels, hidden_channels_1, 5)
[docs]
self.pool = nn.MaxPool2d(2, 2)
[docs]
self.conv2 = nn.Conv2d(hidden_channels_1, hidden_channels_2, 5)
[docs]
self.fc1 = nn.Linear(hidden_channels_2 * pool2_end_h * pool2_end_w, 120)
[docs]
self.fc2 = nn.Linear(120, 84)
[docs]
self.fc3 = nn.Linear(84, self.config["model"]["HyraxCNN"]["output_classes"])
[docs]
def conv2d_output_size(self, input_size, kernel_size, padding=0, stride=1, dilation=1) -> int:
# From https://pytorch.org/docs/stable/generated/torch.nn.Conv2d.html
numerator = input_size + 2 * padding - dilation * (kernel_size - 1) - 1
return int((numerator / stride) + 1)
[docs]
def pool2d_output_size(self, input_size, kernel_size, stride, padding=0, dilation=1) -> int:
# From https://pytorch.org/docs/stable/generated/torch.nn.MaxPool2d.html
numerator = input_size + 2 * padding - dilation * (kernel_size - 1) - 1
return int((numerator / stride) + 1)
[docs]
def forward(self, x):
x, _ = x # Unpack data and ignore labels
x = self.pool(F.relu(self.conv1(x)))
x = self.pool(F.relu(self.conv2(x)))
x = torch.flatten(x, 1)
x = F.relu(self.fc1(x))
x = F.relu(self.fc2(x))
x = self.fc3(x)
return x
[docs]
def train_step(self, batch):
"""This function contains the logic for a single training step. i.e. the
contents of the inner loop of a ML training process.
Parameters
----------
batch : tuple
A tuple containing the inputs and labels for the current batch.
Returns
-------
Current loss value : dict
Dictionary containing the loss value for the current batch.
"""
_, labels = batch
self.optimizer.zero_grad()
outputs = self(batch)
loss = self.criterion(outputs, labels)
loss.backward()
self.optimizer.step()
return {"loss": loss.item()}
@staticmethod
[docs]
def to_tensor(data_dict) -> tuple:
"""Does NOT convert to PyTorch Tensors.
This works exclusively with numpy data types and returns
a tuple of numpy data types."""
import numpy as np
if "data" not in data_dict:
raise RuntimeError("Unable to find `data` key in data_dict")
data = data_dict["data"]
image = np.asarray(data["image"], dtype=np.float32)
label = np.asarray(data.get("label", []), dtype=np.int64)
return (image, label)