Humming

license

Apache-2.0(OSI)

source

public

core features withheld

no

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://api.github.com/repos/inclusionAI/humming recorded 2026-09-02

  Repo metadata for humming's canonical repository - private false, archived false, fork false.

• https://github.com/inclusionAI/humming/blob/main/LICENSE recorded 2026-10-07

  LICENSE body carries the verbatim Apache License 2.0 text under an Ant Group / inclusionAI copyright header.

• https://github.com/inclusionAI/humming/blob/main/README.md recorded 2026-10-07

  README describes Humming as "a high-performance, lightweight, and highly flexible JIT (Just-In-Time) compiled GEMM kernel library specifically designed for quantized inference", supporting "both **Dense GEMM** and **MoE GEMM**" across FP16, BF16, FP8, FP4, INT8 and INT4 activations on NVIDIA GPUs from Turing onward; it names no paid tier, enterprise edition or license-gated build beside the published source.

Humming has no registry package and is built from source, so GitHub stars are the only adoption signal available. A star is not a use.

• https://api.github.com/repos/inclusionAI/humming recorded 2026-09-02

  Repo metadata - stargazers_count = 219 - for inclusionAI/humming.

Humming's JIT-compiled GEMM kernels cover dense and MoE GEMM across six activation/weight dtype combinations and every NVIDIA architecture from Turing onward. That is broad coverage of one optimization family rather than a single-purpose kernel, which places it two steps below Apache TVM's wider compilation scope.

• https://github.com/inclusionAI/humming/blob/main/README.md recorded 2026-10-07

  README describes Humming as "a high-performance, lightweight, and highly flexible JIT (Just-In-Time) compiled GEMM kernel library specifically designed for quantized inference", supporting "both **Dense GEMM** and **MoE GEMM**" across FP16, BF16, FP8, FP4, INT8 and INT4 activations on NVIDIA GPUs from Turing onward; it names no paid tier, enterprise edition or license-gated build beside the published source.