tqkit 0.2.0

Unified toolkit for benchmarking and integrating TurboQuant+ KV-cache compression across inference engines (llama.cpp, vLLM, MLX).

tqkit

Unified toolkit for benchmarking and integrating TurboQuant+ KV-cache compression across LLM inference engines.

What this is

tqkit is a single CLI and Python package that talks to every inference engine that ships TurboQuant+ KV-cache compression:

• llama.cpp (TheTom/llama.cpp@feature/turboquant-kv-cache)
• vLLM (CUDA) (TheTom/vllm@feature/turboquant-kv-cache)
• vLLM (AMD ROCm) (TheTom/vllm@feature/turboquant-amd)
• MLX-Swift (TheTom/mlx@feature/turboquant-plus)
• vllm-swift plugin

You bring the inference engine. tqkit autodetects what's installed, runs the canonical benchmark, and prints a reproducible KV-savings table.

Why this exists

KV cache is the dominant memory cost at long context. TurboQuant+ asymmetric (K=q8_0, V=turbo4) shrinks it ~62% with negligible accuracy loss. The savings replicate across engines and hardware vendors. tqkit is the proof, the tool, and the install path.

For a 14B model at 1M tokens of context:

layout	KV cache size	fits on MI300X 192GB?
FP16	192 GB	no (after weights, ~28 GB free)
TQ+ asym (K=q8_0, V=turbo4)	72 GB	yes

You can verify the math yourself:

pip install tqkit
tq report --model qwen2.5-14b-instruct-1m --ctx 1M --layout tq+asym
tq table --model qwen2.5-14b-instruct-1m

Install

pip install tqkit

Usage

tq backends                                            # autodetect installed engines
tq report --model qwen2.5-14b-instruct-1m --ctx 32K    # KV cache size for one config
tq table --model qwen2.5-14b-instruct-1m               # full layout × ctx grid
tq integrate <backend>                                 # install + serve recipe
tq bench                                               # canonical benchmark (v0.3.0)

Example output:

$ tq report --model qwen2.5-14b-instruct-1m --ctx 1M --layout tq+asym
[KV cache] model: Qwen/Qwen2.5-14B-Instruct-1M
[KV cache] arch: layers=48 kv_heads=8 head_dim=128
[KV cache] layout: tq+asym
[KV cache] per-token: 72.0 KB (vs 192.0 KB FP16)
[KV cache] total @ 1M ctx: 72.0 GB (vs 192.0 GB FP16, 62.5% savings)

Status

v0.2.0 — alpha. KV math + reporter + table work. Canonical bench runner with engine bridges lands in v0.3.0.

License

Apache 2.0.