Add normalized Qwen3.6-27B architecture provenance

#3
by racerxdl - opened

Migrates the Qwen/Qwen3.6-27B adapter to Miru's normalized architecture fingerprint.

Provenance update

  • Model commit: 6a9e13bd6fc8f0983b9b99948120bc37f49c13e9
  • Architecture SHA-256: fa5cf5c9a5cb2215d356e5a1f326c6dd583f0e3e33c55598f692ac94f5a70ae7
  • Architecture hash kind: miru-semantic-v1
  • Hash computed from the preserved snapshot using the original fitting environment's Transformers 5.13.0 and Miru's canonical model_architecture_sha256 algorithm.
  • Removed obsolete model_config_sha256 and model_config_sha256_kind fields rather than replacing their values with the architecture digest.
  • Retained the exact model commit and tokenizer fingerprint.

The adapter was loaded and saved through Miru Tracer's standard lens codec.

  • Adapter SHA-256: 5e5e2686f49b1fd1d74edbcc5a4aff6480dd1d4de2ee33056cd4889b4cb56b80
  • Adapter structure: 63 float16 matrices, each 5120 x 5120

Contributor Public-Domain Certification

I certify that I created this contribution or otherwise have the authority
to submit it. To the extent that I own copyright or related rights in the
adapter and its accompanying metadata, I permanently dedicate those rights to
the public domain under the Unlicense. Where a public-domain dedication is
not legally recognized, I make the contribution available under all
permissions and disclaimers stated by the Unlicense. I have disclosed the
base model and calibration sources, and I am not knowingly submitting
material that I lack permission to distribute. If I am contributing as part
of my employment or for another organization, I certify that I am authorized
to make this dedication on its behalf.

Signed-off-by: Lucas Teske (@racerxdl ), 2026-07-26

rlaneth changed pull request status to merged

Sign up or log in to comment