HazyResearch / train-tk
train with kittens!
☆49Updated 3 weeks ago
Related projects ⓘ
Alternatives and complementary repositories for train-tk
- Experiment of using Tangent to autodiff triton☆72Updated 9 months ago
- Simple and efficient pytorch-native transformer training and inference (batched)☆61Updated 7 months ago
- ☆73Updated 4 months ago
- ☆57Updated 11 months ago
- A MAD laboratory to improve AI architecture designs 🧪☆95Updated 6 months ago
- ☆77Updated 5 months ago
- Make triton easier☆41Updated 5 months ago
- ☆53Updated 10 months ago
- Tree Attention: Topology-aware Decoding for Long-Context Attention on GPU clusters☆104Updated last month
- Experiments for efforts to train a new and improved t5☆76Updated 7 months ago
- ☆55Updated last month
- seqax = sequence modeling + JAX☆133Updated 4 months ago
- The simplest, fastest repository for training/finetuning medium-sized GPTs.☆84Updated this week
- gzip Predicts Data-dependent Scaling Laws☆32Updated 5 months ago
- Triton Implementation of HyperAttention Algorithm☆46Updated 11 months ago
- Latent Diffusion Language Models☆67Updated last year
- A place to store reusable transformer components of my own creation or found on the interwebs☆44Updated 2 weeks ago
- ☆27Updated 4 months ago
- ☆13Updated last month
- Muon optimizer for neural networks: >30% extra sample efficiency, <3% wallclock overhead☆113Updated this week
- Using FlexAttention to compute attention with different masking patterns☆40Updated last month
- ☆18Updated last month
- The source code of our work "Prepacking: A Simple Method for Fast Prefilling and Increased Throughput in Large Language Models"☆56Updated last month
- ☆49Updated 8 months ago
- Proof-of-concept of global switching between numpy/jax/pytorch in a library.☆18Updated 5 months ago
- Collection of autoregressive model implementation☆67Updated this week
- ☆50Updated 6 months ago
- ☆20Updated last year
- Minimal (400 LOC) implementation Maximum (multi-node, FSDP) GPT training☆113Updated 7 months ago
- Cold Compress is a hackable, lightweight, and open-source toolkit for creating and benchmarking cache compression methods built on top of…☆87Updated 3 months ago