LLM Compressor
vLLM (UC Berkeley / vLLM team)LLM Compressor applies quantization and sparsification to large language models for deployment with vLLM, covering weight, activation, KV-cache and attention quantization along with transforms such as SmoothQuant and GPTQ. It reads and writes Hugging Face models, saves in the compressed-tensors format vLLM loads, and supports distributed and disk-offloaded compression of models too large for one device.
Verified 2026-08-18 via the GitHub API and the LICENSE body.
Openness
5 high confidence- license
- Apache-2.0(OSI)
- source
- public
- core-gated
- ungated
Apache-2.0 license body confirmed. The repository is public and unarchived and builds the whole product, and the README describes no paid tier, enterprise edition or license-gated build beside it, so source is public and the core ungated.
- https://github.com/vllm-project/llm-compressor/blob/main/LICENSE recorded 2026-08-18
LICENSE file is the verbatim Apache-2.0 text
- https://api.github.com/repos/vllm-project/llm-compressor recorded 2026-08-18
Repo metadata - license spdx_id Apache-2.0, private false, archived false, default branch main - for vllm-project/llm-compressor.
- https://github.com/vllm-project/llm-compressor/blob/main/README.md recorded 2026-08-18
README describes weight, activation, KV-cache and attention quantization plus sparsification transforms over Hugging Face models, saved in the compressed-tensors format vLLM loads, with DDP and disk offloading for very large models, with no paid tier, enterprise edition or license-key-gated build beside it.
Adoption
3 high confidence256,030 PyPI downloads of `llmcompressor` in the trailing 30 days, which lands in the 100K-1M band of the software usage scale, level 3.
- https://pypistats.org/api/packages/llmcompressor/recent recorded 2026-08-18
last_month downloads = 256,030 for llmcompressor
Capability
3 medium confidenceBanded on the category feature matrix as one optimization family, applied broadly. Placed two bands below the apache-tvm anchor, on a matrix that bands on how much of the model-to-hardware transformation pipeline a product performs, over how many inputs and targets.
- https://github.com/vllm-project/llm-compressor/blob/main/README.md recorded 2026-08-18
README still documents weight, activation, KV-cache and attention quantization plus sparsification transforms over Hugging Face models, saved in the compressed-tensors format vLLM loads, with DDP and disk offloading for very large models.
Verified 2026-08-18