| # Attribution and license scope |
|
|
| PhaGen |
| Copyright 2026 The PhaGen authors |
|
|
| Licensed under the Apache License, Version 2.0 (see `LICENSE`). |
|
|
| ## Third-party code in this repository |
|
|
| ### `modeling_phagen.py` |
| |
| Adapts the MEGABYTE implementation by Phil Wang |
| (https://github.com/lucidrains/MEGABYTE-pytorch). The multi-stage hierarchical |
| transformer, its helper utilities and the per-stage start tokens come from that |
| implementation. The DNA configuration, the masked-diffusion integration and the |
| Transformers model class are added here. |
| |
| > MIT License |
| > |
| > Copyright (c) 2023 Phil Wang |
| > |
| > Permission is hereby granted, free of charge, to any person obtaining a copy |
| > of this software and associated documentation files (the "Software"), to deal |
| > in the Software without restriction, including without limitation the rights |
| > to use, copy, modify, merge, publish, distribute, sublicense, and/or sell |
| > copies of the Software, and to permit persons to whom the Software is |
| > furnished to do so, subject to the following conditions: |
| > |
| > The above copyright notice and this permission notice shall be included in all |
| > copies or substantial portions of the Software. |
| > |
| > THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR |
| > IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
| > FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE |
| > AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER |
| > LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, |
| > OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE |
| > SOFTWARE. |
| |
| ### `generation_utils.py` |
|
|
| Adapts the masked-diffusion generation utilities from Open-dLLM |
| (https://github.com/pengzhangzhi/Open-dLLM), licensed under the Apache License, |
| Version 2.0. The mixin in that file records further lineage from the Dream model's |
| generation utilities. |
|
|
| ## Training framework |
|
|
| The paper model was trained with VeOmni (https://github.com/ByteDance-Seed/VeOmni, |
| Apache-2.0). The training framework is not part of this model repository. |
|
|
| ## Scope of the Apache-2.0 grant |
|
|
| `LICENSE` covers the code and configuration authored for this repository, and applies |
| alongside the upstream terms above for the adapted files. It makes no license claim over |
| the biological datasets used for training or evaluation, which keep their own terms, and |
| it does not relicense any third-party source. |
|
|
| ## Weights |
|
|
| The weights are the frozen paper artifact, released for research use together with the |
| code in this repository. |
|
|
