RobertCsordas/modules
The official repository for our paper "Are Neural Nets Modular? Inspecting Functional Modularity Through Differentiable Weight Masks". We develop a method for analyzing emerging functional modularity in neural networks based on differentiable weight masks and use it to point out important issues in current-day neural networks.
This project helps machine learning researchers understand how different parts of a neural network specialize in certain tasks. It takes a trained neural network and applies differentiable weight masks to reveal which internal components are functionally modular. The output is a series of plots and analysis that visualizes the network's modularity, helping researchers improve network design.
No commits in the last 6 months.
Use this if you are a machine learning researcher studying neural network architectures and want to analyze the functional modularity of your models.
Not ideal if you are looking for a tool to train neural networks or optimize their performance, as this project focuses on post-training analysis.
Stars
46
Forks
9
Language
Python
License
BSD-3-Clause
Category
Last pushed
Oct 03, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/RobertCsordas/modules"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
galilai-group/stable-pretraining
Reliable, minimal and scalable library for pretraining foundation and world models
CognitiveAISystems/MAPF-GPT
[AAAI-2025] This repository contains MAPF-GPT, a deep learning-based model for solving MAPF...
UKPLab/gpl
Powerful unsupervised domain adaptation method for dense retrieval. Requires only unlabeled...
larslorch/avici
Amortized Inference for Causal Structure Learning, NeurIPS 2022
svdrecbd/mhc-mlx
MLX + Metal implementation of mHC: Manifold-Constrained Hyper-Connections by DeepSeek-AI.