tfjgeorge / nngeometry
{KFAC,EKFAC,Diagonal,Implicit} Fisher Matrices and finite width NTKs in PyTorch
☆212Updated this week
Alternatives and similar repositories for nngeometry:
Users that are interested in nngeometry are comparing it to the libraries listed below
- Hessian spectral density estimation in TF and Jax☆123Updated 4 years ago
- PyTorch linear operators for curvature matrices (Hessian, Fisher/GGN, KFAC, ...)☆36Updated last week
- Repository containing Pytorch code for EKFAC and K-FAC perconditioners.☆142Updated last year
- Pytorch implementation of KFAC and E-KFAC (Natural Gradient).☆132Updated 5 years ago
- Efficient PyTorch Hessian eigendecomposition tools!☆371Updated last year
- ASDL: Automatic Second-order Differentiation Library for PyTorch☆185Updated 4 months ago
- Code for the paper: "Tensor Programs II: Neural Tangent Kernel for Any Architecture"☆106Updated 4 years ago
- PyTorch-SSO: Scalable Second-Order methods in PyTorch☆145Updated last year
- ☆66Updated 6 years ago
- NTK reading group☆88Updated 5 years ago
- BackPACK - a backpropagation package built on top of PyTorch which efficiently computes quantities other than the gradient.☆578Updated 3 months ago
- PyTorch implementation of FIM and empirical FIM☆60Updated 6 years ago
- Mode Connectivity and Fast Geometric Ensembles in PyTorch☆270Updated 2 years ago
- ☆241Updated 2 years ago
- Code for experiments in my blog post on the Neural Tangent Kernel: https://eigentales.com/NTK☆172Updated 5 years ago
- ☆157Updated 2 years ago
- ☆67Updated 4 months ago
- paper lists and information on mean-field theory of deep learning☆74Updated 6 years ago
- ☆122Updated 10 months ago
- Collection of algorithms for approximating Fisher Information Matrix for Natural Gradient (and second order method in general)☆139Updated 5 years ago
- Project site for "Your Classifier is Secretly an Energy-Based Model and You Should Treat it Like One"☆422Updated 2 years ago
- Create animations for the optimization trajectory of neural nets☆153Updated last year
- Code for "The Intrinsic Dimension of Images and Its Impact on Learning" - ICLR 2021 Spotlight https://openreview.net/forum?id=XJk19XzGq2J☆68Updated last year
- Code for the article "What if Neural Networks had SVDs?", to be presented as a spotlight paper at NeurIPS 2020.☆75Updated 9 months ago
- Reparameterize your PyTorch modules☆71Updated 4 years ago
- Convolutional Neural Tangent Kernel☆111Updated 5 years ago
- Neural Tangent Kernel Papers☆108Updated 3 months ago
- Rethinking Bias-Variance Trade-off for Generalization of Neural Networks☆49Updated 4 years ago
- Approximating neural network loss landscapes in low-dimensional parameter subspaces for PyTorch☆326Updated last year
- Package for working with hypernetworks in PyTorch.☆123Updated last year