Skip to content

After installing openretina (see installation), we will now showcase a few features of the library by loading a pre-trained model, using it to predict neural activity, and visualizing the model's internal weights.

Loading a Pre-trained Model

openretina provides several pre-trained models. Here's how to load one:

import torch
from openretina.models import load_core_readout_from_remote

# Load a pre-trained model (will download if not already cached)
model = load_core_readout_from_remote("hoefling_2024_base_low_res", "cpu")

Running Inference

We can now use this model by first creating a random stimulus and predicting the activity of each modeled neuron in response.

# Get the appropriate input shape for this model
batch_size = 1
time_steps = 50
input_shape = model.stimulus_shape(time_steps=time_steps, num_batches=batch_size)

# Create random input (in a real scenario, this would be your visual stimulus)
random_stimulus = torch.rand(input_shape)

# Run inference
responses = model.forward(random_stimulus)
print(f"Model predicted responses shape: {responses.shape}")

Visualizing Model Components

The models in openretina use a Core+Readout architecture. The Core, which is shared across all sessions and cells, processes the input stimulus and extracts a rich feature representation of the visual input. The Readout, which is specific to each experimental session, then maps the Core's output feature space to individual neurons' firing-rate predictions for a given stimulus.

We can visualize these components individually, for example by looking at a particular weight in the first layer of the Core. 

# Visualise a specific channel weight for the first convolutional layer
conv_layer_figure = model.core.plot_weight_visualization(layer=0, in_channel=1, out_channel=0)
conv_layer_figure.show()

And by visualizing the weights of the readout layer that predict the activity of an individual neuron based on the features of the Core:

# Visualise the readout weights of a neuron for a particular readout session
session_key = model.readout.readout_keys()[0]
readout_figure = model.readout[session_key].plot_weight_for_neuron(5)
readout_figure.show()

Next Steps

Now that you're familiar with the basics, you can continue to learn about: