ExLlamaV2: Fast inference library for local LLMs (GitHub: turboderp-org/exllamav2)

Fields

id

sources/qhw-src-exllamav2-repo

URL

https://github.com/turboderp-org/exllamav2

Publisher

turboderp-org

Kind

repo

Accessed

2026-10-04

notes

ExLlamaV2 GitHub repository. Archived, development moved to ExLlamaV3. Contains performance benchmarks for GPTQ and EXL2 quantized models on 3090 Ti and 4090 GPUs. Supports paged attention via Flash Attention 2.5.7+. Dynamic generator with batching, prompt caching, and KV cache deduplication.

Notes

ExLlamaV2 GitHub repository. Archived, development moved to ExLlamaV3. Contains performance benchmarks for GPTQ and EXL2 quantized models on 3090 Ti and 4090 GPUs. Supports paged attention via Flash Attention 2.5.7+. Dynamic generator with batching, prompt caching, and KV cache deduplication.

Cross-references

This record references no other records.