ironjr / grokfastLinks
Official repository for the paper "Grokfast: Accelerated Grokking by Amplifying Slow Gradients"
☆560Updated 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☆256Updated 3 weeks ago
- Annotated version of the Mamba paper☆488Updated last year
- The AdEMAMix Optimizer: Better, Faster, Older.☆185Updated 11 months ago
- Quick implementation of nGPT, learning entirely on the hypersphere, from NvidiaAI☆289Updated 2 months ago
- [ICLR 2025] Samba: Simple Hybrid State Space Models for Efficient Unlimited Context Language Modeling☆904Updated 3 months ago
- ☆101Updated last month
- Official JAX implementation of xLSTM including fast and efficient training and inference code. 7B model available at https://huggingface.…☆101Updated 7 months ago
- Open weights language model from Google DeepMind, based on Griffin.☆647Updated 2 months ago
- Repo for "Monarch Mixer: A Simple Sub-Quadratic GEMM-Based Architecture"☆559Updated 8 months ago
- Getting crystal-like representations with harmonic loss☆194Updated 4 months ago
- Minimalistic, extremely fast, and hackable researcher's toolbench for GPT models in 307 lines of code. Reaches <3.8 validation loss on wi…☆349Updated last year
- ☆207Updated 8 months ago
- Implementation of Diffusion Transformer (DiT) in JAX☆291Updated last year
- A repository for log-time feedforward networks☆223Updated last year
- ☆275Updated last year
- Reasoning Computers. Lambda Calculus, Fully Differentiable. Also Neural Stacks, Queues, Arrays, Lists, Trees, and Latches.☆270Updated 9 months ago
- Pretraining and inference code for a large-scale depth-recurrent language model☆816Updated last month
- For optimization algorithm research and development.☆530Updated this week
- Simple, minimal implementation of the Mamba SSM in one pytorch file. Using logcumsumexp (Heisen sequence).☆121Updated 10 months ago
- Official Implementation of "ADOPT: Modified Adam Can Converge with Any β2 with the Optimal Rate"☆428Updated 8 months ago
- Training small GPT-2 style models using Kolmogorov-Arnold networks.☆121Updated last year
- Reverse Engineering the Abstraction and Reasoning Corpus☆295Updated 6 months ago
- CIFAR-10 speedruns: 94% in 2.6 seconds and 96% in 27 seconds☆284Updated last month
- The history files when recording human interaction while solving ARC tasks☆114Updated last week
- ☆118Updated 8 months ago
- The repository for the code of the UltraFastBERT paper☆517Updated last year
- Code for Adam-mini: Use Fewer Learning Rates To Gain More https://arxiv.org/abs/2406.16793☆433Updated 3 months ago
- 🧱 Modula software package☆225Updated last week
- Normalized Transformer (nGPT)☆186Updated 9 months ago
- Code repository for Black Mamba☆254Updated last year