jxbz / modula
Scalable neural net training via automatic normalization in the modular norm.
☆121Updated 3 months ago
Related projects ⓘ
Alternatives and complementary repositories for modula
- ☆128Updated this week
- LoRA for arbitrary JAX models and functions☆132Updated 8 months ago
- ☆73Updated 4 months ago
- A simple library for scaling up JAX programs☆127Updated 2 weeks ago
- A MAD laboratory to improve AI architecture designs 🧪☆95Updated 6 months ago
- Efficient optimizers☆79Updated this week
- ☆197Updated 4 months ago
- Named Tensors for Legible Deep Learning in JAX☆153Updated this week
- Minimal (400 LOC) implementation Maximum (multi-node, FSDP) GPT training☆113Updated 7 months ago
- WIP☆89Updated 3 months ago
- Experiment of using Tangent to autodiff triton☆72Updated 9 months ago
- ☆53Updated 10 months ago
- seqax = sequence modeling + JAX☆133Updated 4 months ago
- If it quacks like a tensor...☆52Updated last week
- ☆40Updated 4 months ago
- Understand and test language model architectures on synthetic tasks.☆162Updated 6 months ago
- Muon optimizer for neural networks: >30% extra sample efficiency, <3% wallclock overhead☆109Updated last week
- Accelerated First Order Parallel Associative Scan☆163Updated 3 months ago
- The simplest, fastest repository for training/finetuning medium-sized GPTs.☆84Updated last week
- Run PyTorch in JAX. 🤝☆200Updated last year
- Normalized Transformer (nGPT)☆66Updated this week
- ☆58Updated 2 years ago
- Multidimensional indexing for tensors☆113Updated last year
- A library for unit scaling in PyTorch☆105Updated 2 weeks ago
- Simple implementation of muP, based on Spectral Condition for Feature Learning. The implementation is SGD only, dont use it for Adam☆69Updated 3 months ago
- ☆46Updated last month
- JAX Synergistic Memory Inspector☆164Updated 4 months ago
- 94% on CIFAR-10 in 2.6 seconds 💨 96% in 27 seconds☆177Updated last week
- Pytorch-like dataloaders in JAX.☆59Updated last month
- Code to reproduce "Transformers Can Do Arithmetic with the Right Embeddings", McLeish et al (NeurIPS 2024)☆178Updated 5 months ago