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.
☆151Aug 9, 2024Updated last year
Alternatives and similar repositories for cold-compress
Users that are interested in cold-compress are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- This repo contains the source code for: Model Tells You What to Discard: Adaptive KV Cache Compression for LLMs☆43Aug 14, 2024Updated last year
- ☆22Apr 17, 2025Updated last year
- Pytorch implementation of our paper accepted by ICML 2024 -- CaM: Cache Merging for Memory-efficient LLMs Inference☆49Jun 19, 2024Updated last year
- ☆14Jun 4, 2024Updated last year
- ☆313Jul 10, 2025Updated 10 months ago
- Deploy open-source AI quickly and easily - Special Bonus Offer • AdRunpod Hub is built for open source. One-click deployment and autoscaling endpoints without provisioning your own infrastructure.
- [ICML 2024] Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference☆384Jul 10, 2025Updated 10 months ago
- [ICLR 2025] Palu: Compressing KV-Cache with Low-Rank Projection☆157Feb 20, 2025Updated last year
- KV cache compression for high-throughput LLM inference☆158Feb 5, 2025Updated last year
- 📰 Must-read papers on KV Cache Compression (constantly updating 🤗).☆705Apr 15, 2026Updated last month
- ☆47Nov 25, 2024Updated last year
- Make triton easier☆50Jun 12, 2024Updated last year
- Keyformer proposes KV Cache reduction through key tokens identification and without the need for fine-tuning☆57Mar 26, 2024Updated 2 years ago
- [ICLR2025] Breaking Throughput-Latency Trade-off for Long Sequences with Speculative Decoding☆148Dec 4, 2024Updated last year
- The Official Implementation of Ada-KV [NeurIPS 2025]☆132Nov 26, 2025Updated 5 months ago
- Serverless GPU API endpoints on Runpod - Get Bonus Credits • AdSkip the infrastructure headaches. Auto-scaling, pay-as-you-go, no-ops approach lets you focus on innovating your application.
- LLM KV cache compression made easy☆1,084May 12, 2026Updated last week
- [NeurIPS'24 Spotlight, ICLR'25, ICML'25] To speed up Long-context LLMs' inference, approximate and dynamic sparse calculate the attention…☆1,212Apr 8, 2026Updated last month
- QAQ: Quality Adaptive Quantization for LLM KV Cache☆55Mar 27, 2024Updated 2 years ago
- ☆38Mar 17, 2025Updated last year
- Boosting 4-bit inference kernels with 2:4 Sparsity☆96Sep 4, 2024Updated last year
- A simple python wrapper for using the Caddy API☆27Apr 28, 2026Updated 3 weeks ago
- GEAR: An Efficient KV Cache Compression Recipefor Near-Lossless Generative Inference of LLM☆183Jul 12, 2024Updated last year
- [ICLR 2025] DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming Heads☆543Feb 10, 2025Updated last year
- ☆164Feb 15, 2025Updated last year
- Deploy to Railway using AI coding agents - Free Credits Offer • AdUse Claude Code, Codex, OpenCode, and more. Autonomous software development now has the infrastructure to match with Railway.
- Easy control for Key-Value Constrained Generative LLM Inference(https://arxiv.org/abs/2402.06262)☆62Feb 13, 2024Updated 2 years ago
- InfiniGen: Efficient Generative Inference of Large Language Models with Dynamic KV Cache Management (OSDI'24)☆187Jul 10, 2024Updated last year
- ☆18Apr 3, 2023Updated 3 years ago
- [NeurIPS'23] H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models.☆518Aug 1, 2024Updated last year
- [ICLR2025] Code and data for paper: Not All Heads Matter: A Head-Level KV Cache Compression Method with Integrated Retrieval and Reasonin…☆42Mar 10, 2025Updated last year
- Fast, Modern, and Low Precision PyTorch Optimizers☆129Dec 29, 2025Updated 4 months ago
- Quantized Attention on GPU☆44Nov 22, 2024Updated last year
- A sample pattern for running CI tests on Modal☆19Apr 12, 2025Updated last year
- Vortex: A Flexible and Efficient Sparse Attention Framework☆53Updated this week
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- [ICML 2024] KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache☆399Nov 20, 2025Updated 5 months ago
- ☆21Mar 3, 2025Updated last year
- ☆234Jun 11, 2024Updated last year
- [ACL 2025] Squeezed Attention: Accelerating Long Prompt LLM Inference☆60Nov 20, 2024Updated last year
- ☆20Jul 12, 2023Updated 2 years ago
- Triton-based implementation of Sparse Mixture of Experts.☆273Oct 3, 2025Updated 7 months ago
- ☆53May 13, 2024Updated 2 years ago