| Solomon v1.1.0 |
| Copyright 2026 Doccy Pty Ltd |
|
|
| Licensed under the Apache License, Version 2.0 (the "License"); you may not use |
| this work except in compliance with the License. You may obtain a copy of the |
| License in the LICENSE file distributed with this work, or at |
|
|
| http://www.apache.org/licenses/LICENSE-2.0 |
|
|
| |
| ATTRIBUTION FOR THE BASE MODEL (Apache License 2.0, section 4(c)) |
| |
|
|
| This work is a DERIVATIVE WORK of: |
|
|
| Qwen/Qwen3.8-27B |
| Copyright 2026 Alibaba Cloud |
| Licensed under the Apache License, Version 2.0 |
| Pinned revision: 1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0 |
| Licence text: licenses/Qwen-Apache-2.0.txt (sha256 bbedc3fda3305820b977265f01b8619d87570a6739de3a5582c3464840f1e57a, 11544 bytes, |
| the exact bytes served at the pinned revision) |
|
|
| The base model weights are NOT redistributed in this repository. They are |
| referenced by the pinned revision above and downloaded by the operator directly |
| from the upstream repository under the upstream licence. |
|
|
| Upstream NOTICE file: NONE. A NOTICE file is absent from Qwen/Qwen3.8-27B at the |
| pinned revision (HTTP 404, retrieved 2026-09-18T03:49Z without credentials). |
| Apache-2.0 section 4(d) therefore imposes no carry-forward obligation here. No |
| upstream attribution text has been invented or paraphrased. |
|
|
| "Qwen" and "Alibaba Cloud" are used nominatively, to identify the base model |
| this adapter was trained against. Apache-2.0 section 6 grants no trademark |
| rights and none are claimed or implied. Nothing here states or implies any |
| endorsement, sponsorship or affiliation. |
|
|
| |
| STATEMENT OF CHANGES (Apache License 2.0, section 4(b)) |
| |
|
|
| No upstream source file is modified, and no upstream file is redistributed |
| except the unmodified licence text at licenses/Qwen-Apache-2.0.txt. |
|
|
| The modification this work carries is a trained LoRA adapter and a set of |
| trained linear answer heads, applied to the base model at inference time: |
|
|
| * LoRA adapter, rank 64, question-side placement, float32 |
| adapter/adapter.safetensors |
| sha256 2addaf841ecc58829ad73081603b8d3e52743c53c6d558a17a1dd30e9bd2dbca |
| 870363376 bytes |
| * Trained linear answer heads |
| adapter/heads.npz |
| sha256 126a9b5487dca937a768a4f228f2d2e7d513900d8ad0e99c4137fbaa42a1aa6f |
| 2053938 bytes |
| * Readout calibration (one positive scalar per task) |
| serving/scope9-readout-temperature-v2.json |
| file sha256 baa7263ca9e865dda230617563e6a615de5050adc620c28924dffef68d527118 |
| payload sha256 682d611ec53c3322905c7e265bf9b9b35a1cd581f6d8b7ea924df864a3d74278 |
|
|
| Training took place between 16 and 20 September 2026. The adapter and the heads |
| are identified by the checksums above. Full change detail is in MODIFICATIONS.md. |
|
|
| |
| THIRD-PARTY CONTENT IN THIS REPOSITORY |
| |
|
|
| The only third-party content distributed here is the unmodified Apache License |
| 2.0 text at licenses/Qwen-Apache-2.0.txt, reproduced to satisfy section 4(a). |
|
|
| No training document, evaluation panel, dataset, corpus, rendered page, cached |
| state, score archive or acceptance fixture is distributed. Every file in this |
| repository was scanned for text originating in third-party source documents |
| before release; see MODIFICATIONS.md, "Third-party text scan". |
|
|
| Everything else in this repository |
| and head weights, the calibration artifact, the serving binding and the |
| documentation |
| Apache-2.0. |
|
|
