AnswerDotAI / cold-compress
Cold Compress is a hackable, lightweight, and open-source toolkit for creating and benchmarking cache compression methods built on top of GPT-Fast, a simple, PyTorch-native generation codebase.
☆128Updated 8 months ago
Alternatives and similar repositories for cold-compress:
Users that are interested in cold-compress are comparing it to the libraries listed below
- Triton-based implementation of Sparse Mixture of Experts.☆210Updated 4 months ago
- ☆122Updated 2 months ago
- ☆197Updated 4 months ago
- Repo for "LoLCATs: On Low-Rank Linearizing of Large Language Models"☆231Updated 2 months ago
- The official repo for "LLoCo: Learning Long Contexts Offline"☆116Updated 10 months ago
- Boosting 4-bit inference kernels with 2:4 Sparsity☆72Updated 7 months ago
- Code for studying the super weight in LLM☆98Updated 4 months ago
- Layer-Condensed KV cache w/ 10 times larger batch size, fewer params and less computation. Dramatic speed up with better task performance…☆148Updated 2 weeks ago
- The simplest implementation of recent Sparse Attention patterns for efficient LLM inference.☆59Updated 3 months ago
- ☆237Updated 11 months ago
- Tree Attention: Topology-aware Decoding for Long-Context Attention on GPU clusters☆126Updated 4 months ago
- GEAR: An Efficient KV Cache Compression Recipefor Near-Lossless Generative Inference of LLM☆159Updated 9 months ago
- Code for exploring Based models from "Simple linear attention language models balance the recall-throughput tradeoff"☆232Updated 2 months ago
- [ICLR2025] Breaking Throughput-Latency Trade-off for Long Sequences with Speculative Decoding☆115Updated 4 months ago
- ☆125Updated last year
- Understand and test language model architectures on synthetic tasks.☆192Updated last month
- The source code of our work "Prepacking: A Simple Method for Fast Prefilling and Increased Throughput in Large Language Models" [AISTATS …☆59Updated 6 months ago
- Explorations into some recent techniques surrounding speculative decoding☆259Updated 4 months ago
- Simple implementation of Speculative Sampling in NumPy for GPT-2.☆93Updated last year
- [NeurIPS 2024] KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantization☆340Updated 8 months ago
- ☆50Updated 5 months ago
- Efficient triton implementation of Native Sparse Attention.☆139Updated 2 weeks ago
- ☆143Updated last year
- [ICML 2024] KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache☆288Updated 3 months ago
- Load compute kernels from the Hub☆115Updated this week
- CUDA and Triton implementations of Flash Attention with SoftmaxN.☆69Updated 10 months ago
- 🔥 A minimal training framework for scaling FLA models☆107Updated 2 weeks ago
- ☆43Updated last year
- Official repository for DistFlashAttn: Distributed Memory-efficient Attention for Long-context LLMs Training☆209Updated 8 months ago
- [ICLR 2025] Palu: Compressing KV-Cache with Low-Rank Projection☆99Updated 2 months ago