Repository navigation
Releases
Releases · chfeng-cs/LMCache
Nightly 2026-11-06 · ROCm 7.2 (gfx942, gfx950)
Nightly ROCm 7.2 wheels built from dev on 2026-11-06,
for AMD Instinct gfx942 (MI300X/MI325X) and gfx950 (MI350X/MI355X),
ABI-matched to the upstream vllm/vllm-openai-rocm image
(torch 2.11, cp312).
Install into an upstream vLLM ROCm container:
pip install lmcache==0.0.0.dev+rocm7.2 --no-deps \
--find-links https://github.com/chfeng-cs/LMCache/releases/expanded_assets/nightly-rocm
Nightly 2026-11-05 · CUDA 12.9
Nightly CUDA 12.9 wheels built from dev on 2026-11-05.
uv pip install lmcache --pre \
--extra-index-url https://download.pytorch.org/whl/cu129 \
--find-links https://github.com/chfeng-cs/LMCache/releases/expanded_assets/nightly-cu129 \
--index-strategy unsafe-best-match
Nightly 2026-11-05 · CUDA 13.0
Nightly CUDA 13.0 wheels built from dev on 2026-11-05.
uv pip install lmcache --pre \
--extra-index-url https://download.pytorch.org/whl/cu130 \
--find-links https://github.com/chfeng-cs/LMCache/releases/expanded_assets/nightly \
--index-strategy unsafe-best-match