yuezhouhu / 2by4-pretrainLinks
Efficient 2:4 sparse training algorithms and implementations
☆55Updated 8 months ago
Alternatives and similar repositories for 2by4-pretrain
Users that are interested in 2by4-pretrain are comparing it to the libraries listed below
Sorting:
- [ICML 2025] XAttention: Block Sparse Attention with Antidiagonal Scoring☆222Updated last month
- A sparse attention kernel supporting mix sparse patterns☆291Updated 6 months ago
- 🔥 A minimal training framework for scaling FLA models☆233Updated 2 weeks ago
- 16-fold memory access reduction with nearly no loss☆104Updated 5 months ago
- [ICLR 2025] COAT: Compressing Optimizer States and Activation for Memory-Efficient FP8 Training☆234Updated 3 weeks ago
- ☆51Updated last year
- ☆140Updated 6 months ago
- Triton implementation of FlashAttention2 that adds Custom Masks.☆133Updated last year
- [ICML 2025] SparseLoRA: Accelerating LLM Fine-Tuning with Contextual Sparsity☆51Updated last month
- ☆124Updated 3 months ago
- Code for paper: [ICLR2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference☆136Updated 3 months ago
- [ICML 2024] Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference☆327Updated last month
- Code for "Accelerating Transformer Pre-training with 2:4 Sparsity"☆24Updated 8 months ago
- HALO: Hadamard-Assisted Low-Precision Optimization and Training method for finetuning LLMs. 🚀 The official implementation of https://arx…☆20Updated 6 months ago
- [ICLR 2025] Palu: Compressing KV-Cache with Low-Rank Projection☆135Updated 6 months ago
- Flash-Muon: An Efficient Implementation of Muon Optimizer☆176Updated 2 months ago
- A Survey of Efficient Attention Methods: Hardware-efficient, Sparse, Compact, and Linear Attention☆172Updated last week
- Efficient triton implementation of Native Sparse Attention.☆209Updated 3 months ago
- ☆21Updated 5 months ago
- PyTorch bindings for CUTLASS grouped GEMM.☆110Updated 3 months ago
- Fast Hadamard transform in CUDA, with a PyTorch interface☆227Updated this week
- SeerAttention: Learning Intrinsic Sparse Attention in Your LLMs☆147Updated 3 weeks ago
- An algorithm for weight-activation quantization (W4A4, W4A8) of LLMs, supporting both static and dynamic quantization☆148Updated 3 months ago
- An efficient implementation of the NSA (Native Sparse Attention) kernel☆114Updated 2 months ago
- GEAR: An Efficient KV Cache Compression Recipefor Near-Lossless Generative Inference of LLM☆165Updated last year
- Best practices for training DeepSeek, Mixtral, Qwen and other MoE models using Megatron Core.☆76Updated last week
- ☆99Updated 2 weeks ago
- Code implementation of GPTAQ (https://arxiv.org/abs/2504.02692)☆59Updated last month
- ☆240Updated 2 months ago
- Patch convolution to avoid large GPU memory usage of Conv2D☆92Updated 7 months ago