ironjr / grokfastLinks
Official repository for the paper "Grokfast: Accelerated Grokking by Amplifying Slow Gradients"
☆569Updated last year
Alternatives and similar repositories for grokfast
Users that are interested in grokfast are comparing it to the libraries listed below
Sorting:
- Efficient optimizers☆279Updated last week
- The AdEMAMix Optimizer: Better, Faster, Older.☆186Updated last year
- Quick implementation of nGPT, learning entirely on the hypersphere, from NvidiaAI☆294Updated 6 months ago
- Annotated version of the Mamba paper☆493Updated last year
- Getting crystal-like representations with harmonic loss☆194Updated 8 months ago
- CIFAR-10 speedruns: 94% in 2.6 seconds and 96% in 27 seconds☆336Updated last month
- 🧱 Modula software package☆316Updated 4 months ago
- A repository for log-time feedforward networks☆224Updated last year
- [ICLR 2025] Samba: Simple Hybrid State Space Models for Efficient Unlimited Context Language Modeling☆936Updated last month
- ☆131Updated last year
- Open weights language model from Google DeepMind, based on Griffin.☆658Updated 6 months ago
- ☆230Updated last year
- Training small GPT-2 style models using Kolmogorov-Arnold networks.☆121Updated last year
- ☆107Updated 5 months ago
- Official JAX implementation of xLSTM including fast and efficient training and inference code. 7B model available at https://huggingface.…☆105Updated 11 months ago
- Reverse Engineering the Abstraction and Reasoning Corpus☆328Updated 10 months ago
- Simple, minimal implementation of the Mamba SSM in one pytorch file. Using logcumsumexp (Heisen sequence).☆128Updated last year
- Minimalistic, extremely fast, and hackable researcher's toolbench for GPT models in 307 lines of code. Reaches <3.8 validation loss on wi…☆355Updated last year
- Repo for "Monarch Mixer: A Simple Sub-Quadratic GEMM-Based Architecture"☆560Updated last year
- Our solution for the arc challenge 2024☆186Updated 6 months ago
- For optimization algorithm research and development.☆555Updated last week
- Reasoning Computers. Lambda Calculus, Fully Differentiable. Also Neural Stacks, Queues, Arrays, Lists, Trees, and Latches.☆283Updated last year
- Official Implementation of "ADOPT: Modified Adam Can Converge with Any β2 with the Optimal Rate"☆432Updated last year
- ☆287Updated last year
- DeMo: Decoupled Momentum Optimization☆198Updated last year
- Implementation of Diffusion Transformer (DiT) in JAX☆300Updated last year
- The boundary of neural network trainability is fractal☆221Updated last year
- Cost aware hyperparameter tuning algorithm☆176Updated last year
- The history files when recording human interaction while solving ARC tasks☆118Updated this week
- Legible, Scalable, Reproducible Foundation Models with Named Tensors and Jax☆687Updated last month