Skip to content

Latest commit

 

History

History
16 lines (13 loc) · 856 Bytes

File metadata and controls

16 lines (13 loc) · 856 Bytes

Third-party notices

This source repository does not contain model weights.

  • The runtime overlay is derived from vLLM and is distributed under the Apache License 2.0; see LICENSE.
  • The assembled model references Qwen/Qwen3.8-Flash-Next and is governed by the Qwen Community License supplied with the upstream checkpoint.
  • Intel's AutoRound checkpoint and RadixArk's NVFP4 checkpoint remain governed by their respective model cards, notices, and inherited Qwen terms.
  • Humming kernels, PyTorch, CUDA, Docker, and Hugging Face tooling are external dependencies and are not redistributed by this source repository.

The checkpoint assembly script copies the upstream model license into the Hugging Face upload tree. Users are responsible for reviewing the current upstream terms before publishing or using an assembled checkpoint.