Zyphra / tree_attention
Tree Attention: Topology-aware Decoding for Long-Context Attention on GPU clusters
☆94Updated 2 weeks ago
Related projects: ⓘ
- Understand and test language model architectures on synthetic tasks.☆156Updated 4 months ago
- Token Omission Via Attention☆118Updated 7 months ago
- Layer-Condensed KV cache w/ 10 times larger batch size, fewer params and less computation. Dramatic speed up with better task performance…☆127Updated 2 weeks ago
- A MAD laboratory to improve AI architecture designs 🧪☆84Updated 4 months ago
- Code for exploring Based models from "Simple linear attention language models balance the recall-throughput tradeoff"☆206Updated last month
- Cold Compress is a hackable, lightweight, and open-source toolkit for creating and benchmarking cache compression methods built on top of…☆73Updated last month
- Minimal (400 LOC) implementation Maximum (multi-node, FSDP) GPT training☆110Updated 5 months ago
- ☆50Updated last month
- ☆53Updated 8 months ago
- The source code of our work "Prepacking: A Simple Method for Fast Prefilling and Increased Throughput in Large Language Models"☆57Updated 5 months ago
- ☆68Updated 2 months ago
- ☆169Updated this week
- ☆66Updated 3 months ago
- Code to reproduce "Transformers Can Do Arithmetic with the Right Embeddings", McLeish et al (2024)☆169Updated 3 months ago
- seqax = sequence modeling + JAX☆129Updated 2 months ago
- Simple and fast low-bit matmul kernels in CUDA☆48Updated this week
- Official Repository of The Mamba in the Llama: Distilling and Accelerating Hybrid Models☆130Updated this week
- ☆48Updated 3 months ago
- Code repository for the c-BTM paper☆105Updated 11 months ago
- PyTorch implementation of models from the Zamba2 series.☆63Updated last month
- ☆129Updated last year
- Pytorch implementation of the PEER block from the paper, Mixture of A Million Experts, by Xu Owen He at Deepmind☆105Updated 3 weeks ago
- Scalable neural net training via automatic normalization in the modular norm.☆108Updated 3 weeks ago
- Triton-based implementation of Sparse Mixture of Experts.☆166Updated 3 weeks ago
- Some preliminary explorations of Mamba's context scaling.☆184Updated 7 months ago
- Language models scale reliably with over-training and on downstream tasks☆91Updated 5 months ago
- ☆75Updated this week
- ☆61Updated 2 months ago
- Experiments for efforts to train a new and improved t5☆76Updated 5 months ago
- 🚀 Efficiently (pre)training foundation models with native PyTorch features, including FSDP for training and SDPA implementation of Flash…☆150Updated last week