ExLlamaV2

infrastructureinferencetiered

A fast inference library for running LLMs locally on modern consumer-class GPUs - turboderp-org/exllamav2

There was an error while loading. Please reload this page. ExLlamaV2 is an inference library for running local LLMs on modern consumer GPUs. The official and recommended backend server for ExLlamaV2 is TabbyAPI, which provides an OpenAI-compatible API for local or remote inference, with extended features like HF model downloading, embedding model support and support for HF Jinja2 chat templates. See the wiki for help getting started. The dynamic generator supports all inference, sampling and speculative decoding features of the previous two generators, consolidated into one API (with the exception of FP8 cache, though the Q4 cache mode is supported and performs better anyway, see here.) The generator is explained in detail here. See the full, updated examples here. Some quick tests to compare performance with ExLlama V1. There may be more performance optimizations in the future, and speeds will vary across GPUs, with slow CPUs still being a potential bottleneck: To install from the repo you'll need the CUDA Toolkit and either gcc on Linux or (Build Tools for) Visual Studio on Windows). Also make sure you have an appropriate version of PyTorch, then run: A simple console chatbot is included. Run it with: To install the current dev version, clone the repo and run the setup script: This will install the "JIT version" of the package, i.e. it will install the Python components without building the C++ extension in the process. Instead, the extension will be built the first time the library is used, then cached in ~/.cache/torch_extensions for subsequent use. Either download an appropriate wheel or install directly from the appropriate URL: A PyPI package is available as well. This is the same as the JIT version (see above). It can be installed with: ExLlamaV2 supports the same 4-bit GPTQ models as V1, but also a new "EXL2" format. EXL2 is based on the same optimization method as GPTQ and supports 2, 3, 4, 5, 6 and 8-bit quantization. The format allows for mixing quantization levels within a model to achieve any average bitrate between 2 and 8 bits per weight. Moreover, it's possible to apply multiple quantization levels to each linear layer, producing something akin to sparse quantization wherein more important weights (columns) are quantized with more bits. The same remapping trick that lets ExLlama work efficiently with act-order models allows this mixing of formats to happen with little to no impact on performance. Parameter selection is done automatically by quantizing each matrix multiple times, measuring the quantization error (with respect to the chosen calibration data) for each of a number of possible settings, per layer. Finally, a combination is chosen that minimizes the maximum quantization error over the entire model while meeting a target average bitrate. In my tests, this scheme allows Llama2 70B to run on a single 24 GB GPU with a 2048-token context, producing coherent and mostly stable output with 2.55 bits per weight

Mentions (30d)

Reviews

Platforms

Sentiment

0 positive

Share:Twitter LinkedIn

Product Screenshots

Features & Use Cases

Features

New generator with dynamic batching, smart prompt caching, K/V cache deduplication and simplified APIUh oh!Method 1: Install from sourceMethod 2: Install from release (with prebuilt extension)Method 3: Install from PyPIConversionEvaluationCommunityHuggingFace reposResources

Company Intel

Industry

information technology & services

Employees

6,000

Funding Stage

Other

Total Funding

$7.9B

Developer Ecosystem

HuggingFace models

Top Mention

twitter@@github1,359 engagement4/6/2026

Web & game developers, this is your jam. Gamedev.js Jam 2026 is back. 🎮 🗓 April 13-26, 2026 🌐 Build an HTML5 game in 13 days 🏆 Prizes + expert feedback 💬 Active community + Discord Theme revealed on day one. Ship something weird. Ship something fun. Just ship it. https://t.co/KPphbM8rTz

ExLlamaV2

Compare ExLlamaV2 With