From 3611a2d8c33db8a0b300c608201f3d8fccd7cfbb Mon Sep 17 00:00:00 2001 From: TheTom Date: Fri, 12 Jun 2026 10:05:49 -0500 Subject: [PATCH] README: add credited community adopters to Run It Today table --- README.md | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/README.md b/README.md index 57549bc..3aff0ee 100644 --- a/README.md +++ b/README.md @@ -20,6 +20,10 @@ This repository is the **research home**: the Python reference implementation, t | [Atlas](https://github.com/Avarok-Cybersecurity/atlas) | Rust, Metal | Integrated | turbo4 KV cache append/decode kernels | | [mlxcel](https://github.com/lablup/mlxcel) | Rust, MLX | Community port | [Turbo KV cache docs](https://github.com/lablup/mlxcel/blob/main/docs/turbo-kv-cache.md). Thanks to [@inureyes](https://github.com/inureyes) and the Lablup team for the careful port and upstream attribution | | [MLX Python fork](https://github.com/TheTom/mlx/tree/feature/turboquant-plus) | Apple Silicon, Python | Experimental | `TurboKVCache` drop-in for mlx-lm and mlx-vlm; see [docs/mlx-port.md](docs/mlx-port.md) | +| [turboquant-pytorch](https://github.com/tonbistudio/turboquant-pytorch) | PyTorch | Independent implementation | From-scratch PyTorch TurboQuant for KV cache; applies this repo's layer-adaptive compression and Sparse V findings | +| [turboquant-vllm](https://github.com/varjoranta/turboquant-vllm) | CUDA, vLLM package | Community package | TQ+ KV cache for vLLM with fused CUDA kernels; builds on this repo's block_size=128 and K-dominance research | +| [Quansloth](https://github.com/PacifAIst/Quansloth) | Local server product | Community product | Air-gapped local AI server built on the TurboQuant+ stack | +| [RotorQuant](https://github.com/scrya-com/rotorquant) | llama.cpp | Comparison codec | Block-diagonal rotation alternative; ships `turbo3`/`turbo4` alongside its own formats for side-by-side comparison | ## Key Findings