DeepSpec: a full-stack codebase for training and evaluating speculative decoding algorithms
☆6,810Jul 9, 2026Updated 3 weeks ago
Alternatives and similar repositories for DeepSpec
Users that are interested in DeepSpec are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- DFlash: Block Diffusion for Flash Speculative Decoding☆5,547May 10, 2026Updated 2 months ago
- Train speculative decoding models effortlessly and port them smoothly to SGLang serving.☆1,027Updated this week
- Official implementation of “Domino: Decoupling Causal Modeling from Autoregressive Drafting in Speculative Decoding”.☆124Updated this week
- slime is an LLM post-training framework for RL Scaling.☆7,696Updated this week
- A kernel library written in tilelang☆1,683Apr 23, 2026Updated 3 months ago
- Managed hosting for WordPress and PHP on Cloudways • AdManaged hosting for WordPress, Magento, Laravel, or PHP apps, on multiple cloud providers. Deploy in minutes on Cloudways by DigitalOcean.
- TokenSpeed is a speed-of-light LLM inference engine.☆1,751Updated this week
- Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels☆7,007Updated this week
- JetSpec: Breaking the Scaling Ceiling of Speculative Decoding with Causal Parallel Tree Drafting☆171Jun 27, 2026Updated last month
- SGLang is a high-performance serving framework for large language models and multimodal models.☆30,926Updated this week
- FlashKDA: high-performance Kimi Delta Attention kernels☆982Updated this week
- FlashInfer: Kernel Library for LLM Serving☆6,062Updated this week
- DeepGEMM: clean and efficient BLAS kernel library on GPU☆7,581Jul 20, 2026Updated last week
- Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.☆6,083Updated this week
- high-performance linear attention kernel library built on TileLang☆619Updated this week
- GPUs on demand by Runpod - Special Offer Available • AdRun AI, ML, and HPC workloads on powerful cloud GPUs—without limits or wasted spend. Deploy GPUs in under a minute and pay by the second.
- Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models☆4,571Jan 14, 2026Updated 6 months ago
- Tile-Based Runtime for Ultra-Low-Latency LLM Inference☆1,600Jul 14, 2026Updated 2 weeks ago
- A high-throughput and memory-efficient inference and serving engine for LLMs☆87,317Updated this week
- LMCache: Supercharge Your LLM with the Fastest KV Cache Layer☆10,938Updated this week
- Accelerating MoE with IO and Tile-aware Optimizations☆733Jul 4, 2026Updated 3 weeks ago
- Nano vLLM☆14,679Apr 26, 2026Updated 3 months ago
- ☆385Updated this week
- A unified library for building, evaluating, and storing speculative decoding algorithms for LLM inference in vLLM☆669Updated this week
- verl/HybridFlow: A Flexible and Efficient RL Post-Training Framework☆22,711Updated this week
- Managed hosting for WordPress and PHP on Cloudways • AdManaged hosting for WordPress, Magento, Laravel, or PHP apps, on multiple cloud providers. Deploy in minutes on Cloudways by DigitalOcean.
- A compact implementation of SGLang, designed to demystify the complexities of modern LLM serving systems.☆4,645May 17, 2026Updated 2 months ago
- 🚀 Efficient implementations for emerging model architectures☆5,473Updated this week
- Distributed Compiler based on Triton for Parallel Systems☆1,503Jul 20, 2026Updated last week
- Official Implementation of EAGLE-1 (ICML'24), EAGLE-2 (EMNLP'24), and EAGLE-3 (NeurIPS'25).☆2,481Feb 20, 2026Updated 5 months ago
- Miles is an enterprise-facing reinforcement learning framework for LLM and VLM post-training, forked from and co-evolving with slime.☆1,809Updated this week
- High Performance LLM Inference Operator Library☆1,070Updated this week
- A framework for efficient model inference with omni-modality models☆5,719Updated this week
- A Quirky Assortment of CuTe Kernels☆1,080Updated this week
- A PyTorch native library for training speculative decoding models☆212Updated this week
- 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.
- DeepSeek 4 Flash and PRO local inference engine for Metal, CUDA and ROCm☆19,409Updated this week
- ☆780Jun 2, 2026Updated last month
- Run GLM-5.2 (744B MoE) on a 25GB-RAM consumer machine — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦☆20,872Updated this week
- A Flexible Framework for Experiencing Heterogeneous LLM Inference/Fine-tune Optimizations☆19,093Updated this week
- mKernel: fast multi-node, multi-GPU fused kernels☆257Updated this week
- A fast communication-overlapping library for tensor/expert parallelism on GPUs.☆1,350Aug 28, 2025Updated 11 months ago
- Unsloth is a local UI for training and running Gemma 4, Qwen3.6, DeepSeek, Kimi, GLM and other models.☆69,060Updated this week