falseresmem 1.0.3

A pip-installable PyTorch model package for False Image Memorability prediction.

FalseResMem Model Package

A pip-installable Python package containing a custom PyTorch model (FalseResMem) for prediction of false alarms in visual memory recognition tasks.

Contact

Author: Anastasiia Mikhailova
Email: [e-mail]

Installation

You can install the package from PyPI with pip:

pip install falseresmem

Usage

Here's a basic example of how to import and use the model after installation:

Low-level (raw tensors):

import torch
from falseresmem import load_model

model = load_model()
model.eval()
input_tensor = torch.randn(1, 3, 224, 224)
output = model(input_tensor)
print(output)

High-level (image file):

from falseresmem import predict_image

prob = predict_image("test.jpg").item()
print(f"False alarm probability: {prob:.3f}")

Citation

Mikhailova, A. & Bainbridge, W. A. (2026). FalseResMem: A Neural Network to Predict False Alarms in Image Memory. Journal of Vision.

Changelog

v1.0.3 (2026‑05‑07)

• Updated license to University of Chicago FalseResMem License.
• Added citation for FalseResMem abstract at VSS.

v1.0.2 (2026‑03‑20)

• Fixed internal package name in resource loading: now uses "falseresmem" instead of "FalseResMem" for model.pt.
• Updated README example to match: from falseresmem import load_model.

v1.0.0 (2026‑03‑20)

• Now includes model weights for predictions, so the package works immediately after installation without requiring manual weight downloads.

v0.1.0 (2026‑03‑11)

• Initial release: package contains prediction code without bundled model weights.