ironjr / grokfastLinks
Official repository for the paper "Grokfast: Accelerated Grokking by Amplifying Slow Gradients"
☆559Updated last year
Alternatives and similar repositories for grokfast
Users that are interested in grokfast are comparing it to the libraries listed below
Sorting:
- The AdEMAMix Optimizer: Better, Faster, Older.☆184Updated 10 months ago
- Efficient optimizers☆252Updated last week
- Quick implementation of nGPT, learning entirely on the hypersphere, from NvidiaAI☆289Updated 2 months ago
- Annotated version of the Mamba paper☆487Updated last year
- [ICLR 2025] Samba: Simple Hybrid State Space Models for Efficient Unlimited Context Language Modeling☆901Updated 3 months ago
- Official JAX implementation of xLSTM including fast and efficient training and inference code. 7B model available at https://huggingface.…☆98Updated 6 months ago
- ☆206Updated 8 months ago
- Implementation of Diffusion Transformer (DiT) in JAX☆286Updated last year
- Simple, minimal implementation of the Mamba SSM in one pytorch file. Using logcumsumexp (Heisen sequence).☆120Updated 9 months ago
- ☆100Updated 2 weeks ago
- For optimization algorithm research and development.☆525Updated this week
- CIFAR-10 speedruns: 94% in 2.6 seconds and 96% in 27 seconds☆274Updated 3 weeks ago
- A repository for log-time feedforward networks☆222Updated last year
- Getting crystal-like representations with harmonic loss☆192Updated 4 months ago
- ☆275Updated last year
- Normalized Transformer (nGPT)☆185Updated 8 months ago
- Open weights language model from Google DeepMind, based on Griffin.☆645Updated 2 months ago
- The boundary of neural network trainability is fractal☆215Updated last year
- Training small GPT-2 style models using Kolmogorov-Arnold networks.☆121Updated last year
- 🧱 Modula software package☆216Updated last week
- Repo for "Monarch Mixer: A Simple Sub-Quadratic GEMM-Based Architecture"☆555Updated 7 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
- The repository for the code of the UltraFastBERT paper☆516Updated last year
- Code for Adam-mini: Use Fewer Learning Rates To Gain More https://arxiv.org/abs/2406.16793☆431Updated 2 months ago
- DeMo: Decoupled Momentum Optimization☆190Updated 8 months ago
- Code to train and evaluate Neural Attention Memory Models to obtain universally-applicable memory systems for transformers.☆318Updated 9 months ago
- Pretraining and inference code for a large-scale depth-recurrent language model☆808Updated 2 weeks ago
- ☆304Updated last year
- Muon is an optimizer for hidden layers in neural networks☆1,390Updated 3 weeks ago
- Accelerate, Optimize performance with streamlined training and serving options with JAX.☆293Updated this week