seedpack 0.1.0__tar.gz

seedpack-0.1.0/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

seedpack-0.1.0/NOTICE

@@ -0,0 +1,6 @@
+SeedPack
+Copyright 2026 NeuroSquad
+
+This product includes the quantisation sources of ggml / llama.cpp
+(seedpack/ggml/), Copyright (c) 2023-2026 The ggml authors, MIT License
+(see seedpack/ggml/LICENSE).

seedpack-0.1.0/PKG-INFO

@@ -0,0 +1,111 @@
+Metadata-Version: 2.4
+Name: seedpack
+Version: 0.1.0
+Summary: Lossless GPU archiver for GGUF model repositories: the BF16 model and all its quants in one archive about the size of a single compressed BF16 file
+Author: NeuroSquad
+License-Expression: Apache-2.0
+Project-URL: Homepage, https://github.com/glmn-ai/seedpack
+Project-URL: Source, https://github.com/glmn-ai/seedpack
+Project-URL: Issues, https://github.com/glmn-ai/seedpack/issues
+Project-URL: Example archive, https://huggingface.co/neurosquad/Qwen3-0.6B-SEEDPACK
+Keywords: gguf,llama.cpp,lossless,compression,archiver,quantization,cuda,llm,huggingface
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Environment :: GPU :: NVIDIA CUDA :: 13
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: Operating System :: Microsoft :: Windows
+Classifier: Operating System :: POSIX :: Linux
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: System :: Archiving :: Compression
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+License-File: NOTICE
+License-File: seedpack/ggml/LICENSE
+Requires-Dist: numpy>=2.0
+Requires-Dist: numba>=0.60
+Requires-Dist: zstandard>=0.23
+Requires-Dist: blake3>=1.0
+Requires-Dist: cupy-cuda13x>=14.0
+Requires-Dist: nvidia-cuda-nvrtc>=13.0
+Requires-Dist: nvidia-cuda-runtime>=13.0
+Requires-Dist: nvidia-nvjitlink>=13.0
+Requires-Dist: huggingface_hub>=0.30
+Dynamic: license-file
+
+<p align="center">
+  <img src="https://raw.githubusercontent.com/glmn-ai/seedpack/main/assets/banner.png" alt="SeedPack" width="860">
+</p>
+
+# SeedPack
+
+**Lossless GPU archiver for GGUF model repositories.** A whole repository, the BF16 model and all its quants from IQ1 to Q8, packs into one archive about the size of a single compressed BF16 file. Every file comes back byte for byte.
+
+| unsloth/Qwen3-0.6B-GGUF, BF16 + 25 quants (10.95 GB) | Archive | Ratio |
+|---|---:|---:|
+| **SeedPack** | **782 MB** | **7.14%** |
+| lmz 1.3.0 | 4.54 GB | 41.47% |
+| zstd 1.5.7 -19 --long=30 | 6.57 GB | 59.99% |
+| 7-Zip 23.01 -mx9 | 10.21 GB | 93.26% |
+| ZipNN 0.5.4 | 10.33 GB | 94.34% |
+
+The archive is smaller than the repository's largest single quant (UD-Q8_K_XL, 844 MB). The full benchmark against 10 other compressors, with speeds and single files, is on the [example archive's page](https://huggingface.co/neurosquad/Qwen3-0.6B-SEEDPACK).
+
+## Install
+
+```bash
+pip install seedpack
+```
+
+Requirements:
+- An NVIDIA GPU, Turing (RTX 20xx) or newer, with a CUDA 13 driver (R580 or newer).
+- Python 3.11+.
+- Windows 10/11 or Linux. Tested on an RTX 5070 under Windows 10 (Python 3.11) and Ubuntu on WSL2 (Python 3.14).
+- No CUDA Toolkit is needed: NVRTC, nvJitLink and the CUDA headers come from pip.
+
+Archives are portable between operating systems: an archive packed on Windows unpacks bit-exactly on Linux, and the other way round. The first run compiles the GPU quantizers for your card (about 15 s) and caches them in `%LOCALAPPDATA%\seedpack\cache` or `~/.cache/seedpack`; set `SEEDPACK_CACHE` to move the cache.
+
+## Use
+
+```bash
+# straight from the Hugging Face Hub: download once, extract what you need
+seedpack list hf://neurosquad/Qwen3-0.6B-SEEDPACK
+seedpack x    hf://neurosquad/Qwen3-0.6B-SEEDPACK Qwen3-0.6B-Q4_K_M.gguf -o models
+seedpack x    hf://neurosquad/Qwen3-0.6B-SEEDPACK "*Q4_K*" "*Q8_0*" -o models
+seedpack x    hf://neurosquad/Qwen3-0.6B-SEEDPACK -o models      # everything
+
+# local archives
+seedpack pack  models/Qwen3-0.6B-GGUF -o qwen3.varc               # files and/or folders
+seedpack x     qwen3.varc Qwen3-0.6B-Q4_K_M.gguf -o out
+seedpack test  qwen3.varc                                         # decode and verify, write nothing
+```
+
+Notes on the commands:
+- `hf://OWNER/REPO[/FILE.varc][@REVISION]` and `https://huggingface.co/...` links work wherever an archive is expected.
+- `list` reads only the archive's index over the network; `x` and `test` download the archive once into the Hugging Face cache.
+- Extracted files are checked against BLAKE3 hashes. Every chunk of the archive carries a CRC32.
+
+## How it works
+
+A GGUF repository ships the same weights 20–30 times: the BF16 model plus each quantization of it. To a general-purpose compressor that looks like gigabytes of almost incompressible bytes. SeedPack uses the fact that every quant is a deterministic function of the same weights:
+
+- **The seed is stored once.** The most precise model (the BF16 file, or the senior quant when there is none) is coded with an entropy coder built for its weights.
+- **Quants are re-derived, not stored.** Each quant is computed again on the GPU by a bit-exact CUDA port of llama.cpp's quantizers (`ggml-quants.c`, compiled without FMA contraction), with the repository's importance matrix.
+- **Only the residual is kept.** Where a re-derived quant differs from the published file, the residual is coded with the quantizer's own probabilities: scales first, then each code relative to its rounding boundary. For Qwen3-0.6B, all 25 quants together cost about 1 MB on top of the seed.
+- **Duplicates are stored once.** Identical tensors in different files are kept once. Files that are not GGUF are compressed with LZMA or zstd.
+
+## Links
+
+- Source and research notes: <https://github.com/glmn-ai/seedpack>
+- Example archive: <https://huggingface.co/neurosquad/Qwen3-0.6B-SEEDPACK>
+- A [NeuroSquad](https://huggingface.co/neurosquad) project.
+
+## License
+
+Apache-2.0. The bundled llama.cpp / ggml quantization sources (`seedpack/ggml/`) are MIT-licensed; see `NOTICE`.

seedpack-0.1.0/README.md

@@ -0,0 +1,100 @@
+<p align="center"><img src="assets/banner.png" alt="SeedPack" width="860"></p>
+
+# SeedPack
+
+Архиватор без потерь для репозиториев GGUF-моделей, проект [NeuroSquad](https://huggingface.co/neurosquad). Весь репозиторий — BF16-модель и все её кванты от IQ1 до Q8 — сжимается в один архив размером примерно с один сжатый BF16-файл. Каждый файл восстанавливается **побайтово**.
+
+Идея: кванты не хранятся, а «вырастают» из BF16, как из семени. Квантизаторы llama.cpp детерминированы, поэтому архив повторяет квантизацию на GPU из BF16 с imatrix репозитория и бит-в-бит совпадающей копией кода `ggml-quants.c`. Хранится только разница с результатом, закодированная вероятностями самого квантизатора.
+
+Готовый архив на Hugging Face: [neurosquad/Qwen3-0.6B-SEEDPACK](https://huggingface.co/neurosquad/Qwen3-0.6B-SEEDPACK) — весь unsloth/Qwen3-0.6B-GGUF (BF16 и 25 квантов, 10.95 ГБ) в одном файле, с бенчмарком против 11 архиваторов.
+
+## Результаты
+
+| Что сжимается | Исходно | Архив | % | lmz 1.3.0 |
+|---|---|---|---|---|
+| unsloth/Qwen3-0.6B-GGUF целиком (BF16 + 25 квантов) | 10.95 ГБ | 781.7 МБ | 7.14% | 41.5% |
+| unsloth/Qwen3.5-0.8B-GGUF целиком | 14.08 ГБ | 1.12 ГБ | 7.96% | 41.2% |
+| unsloth/Qwen3.5-2B-GGUF целиком | 34.57 ГБ | 2.91 ГБ | 8.41% | 43.7% |
+| только кванты Qwen3.5-2B (21 файл) + imatrix | 28.13 ГБ | 2.55 ГБ | 9.06% | 40.2% |
+| только кванты Qwen3.5-2B без imatrix | 28.13 ГБ | 3.02 ГБ | 10.74% | 40.2% |
+| один Q8_0 (Qwen3-0.6B) | 639 МБ | 576 МБ | 90.1% | 93.0% |
+| один Q4_K_M (Qwen3-0.6B) | 397 МБ | 372 МБ | 93.8% | 95.0% |
+| один BF16 (Qwen3.5-2B) | 3.78 ГБ | 2.46 ГБ | 65.2% | 65.9% |
+
+Скорость, время и все наборы — в `bench_competitors/VS_LMZ.md`; план развития — в `ROADMAP.md`.
+Распаковка проверяет контрольные суммы. Измерено на RTX 5070, Core i7-8700K (6 ядер, 12 потоков) и NVMe-диске.
+
+## Установка
+
+```
+pip install seedpack
+```
+
+Нужны видеокарта NVIDIA (Turing и новее) с драйвером CUDA 13 (R580+) и Python 3.11+; CUDA Toolkit не нужен (NVRTC, nvJitLink и заголовки ставятся из pip). Проверено на RTX 5070: Windows 10 (Python 3.11) и Ubuntu в WSL2 (Python 3.14). Архивы переносимы между ОС: собранный на Windows распаковывается на Linux бит-в-бит и наоборот. При первом запуске квантизаторы компилируются под вашу видеокарту (около 15 с) и кэшируются в `%LOCALAPPDATA%\seedpack\cache` (или `~/.cache/seedpack`, или `SEEDPACK_CACHE`).
+
+## Использование
+
+```
+seedpack pack   ВХОДЫ... -o архив.varc          # файлы и/или папки
+seedpack unpack АРХИВ [ИМЕНА|МАСКИ] -o DIR       # всё или выбранные файлы (alias: x)
+seedpack list   АРХИВ
+seedpack test   АРХИВ [ИМЕНА|МАСКИ]              # распаковка в памяти + проверка
+```
+
+АРХИВ — файл `.varc` или архив на Hugging Face: `hf://OWNER/REPO[/FILE.varc][@REVISION]` либо ссылка `https://huggingface.co/...`. `list` читает по сети только индекс, `unpack` и `test` один раз скачивают архив в кэш Hugging Face.
+
+Примеры:
+```
+seedpack x hf://neurosquad/Qwen3-0.6B-SEEDPACK Qwen3-0.6B-Q4_K_M.gguf -o models
+seedpack pack models/Qwen3-0.6B-GGUF -o qwen3.varc
+seedpack x qwen3.varc "*IQ1*" README.md -o out
+```
+
+Роли файлов определяются автоматически:
+- GGUF, где больше половины весов в BF16, — это база (модель, mmproj);
+- остальные GGUF выводятся из баз; кванты без BF16-модели — из старшего кванта (семени, `seedpack/seed.py`);
+- одинаковые тензоры разных файлов хранятся один раз;
+- прочие файлы сжимаются LZMA или zstd.
+
+Каждый файл проверяется по BLAKE3 (если пакета нет — по SHA-256), каждая часть архива — по CRC32.
+
+## Разработка
+
+```
+git clone https://github.com/glmn-ai/seedpack && cd seedpack
+python -m venv .venv
+.venv\Scripts\pip install -r requirements.txt -e .
+sh seedpack/ggml/build.sh         # CPU-библиотека квантизаторов (MinGW gcc): нужна только исследовательским скриптам
+python varc.py --help             # то же, что seedpack
+```
+
+Таблицы IQ-сеток (`seedpack/ggml/iqtables.npz`) построены CPU-библиотекой и лежат в пакете; без файла они строятся заново (около 25 с, нужна `ggmlq.dll`).
+
+Публикация: `python -m build` → `twine upload dist/*`.
+
+## Устройство
+
+- `seedpack/cli.py` — CLI (`seedpack`, из исходников — `python varc.py`); `seedpack/container.py` — формат `.varc` (части + индекс в конце, произвольный доступ).
+- `seedpack/repo3.py` — ядро: M28/M29 (базы, предсказание квантов, остатки, конвейер).
+- `seedpack/gpuquant.py` — генерация CUDA-кода из `ggml-quants.c`, бит-в-бит с CPU; варианты квантизаторов старых версий llama.cpp.
+- `seedpack/ptxfix.py` — обход ошибки ptxas при делении на константу.
+- `seedpack/bf16v2.py` — кодек BF16 (R2); `seedpack/gpu_bf16.py` — M24.
+- `seedpack/gpu3.py` (на основе `seedpack/gpu2.py`) — M21/M22: кодек Q8_0 без базы (rANS с контекстами, предсказание строк по похожим строкам).
+- `seedpack/seed.py` — семя: старший квант набора без BF16 и его кодек.
+- `seedpack/qr.py` — QR: GPU-кодирование остатка вероятностями квантизатора (масштабы, затем коды), и тот же кодер без семени для одиночных файлов.
+- `seedpack/requant.py` — CPU-квантизаторы через `ggmlq.dll`, загрузка imatrix (legacy `.dat` и GGUF).
+- `bench*.py`, `results*.json` — замеры всех методов M1–M29; `research/` — исследования кодеков.
+- `bench_competitors/` — сравнение с другими архиваторами: `run_bench.py`, `quant_sets.py`, `vs_lmz.py`; `hf_bench.py` и `hf_card.py` — бенчмарк и карточка релиза Qwen3-0.6B на Hugging Face.
+- `assets/` — логотип (`logo.svg`) и баннер (`banner.png`, исходник `banner.html`).
+- `seedpack/ggml/` — исходники квантизации llama.cpp (MIT, см. `seedpack/ggml/LICENSE`).
+
+## Ограничения
+
+- Нужна видеокарта NVIDIA.
+- Кванты без BF16-модели: семенем становится старший квант (больше всего бит на вес: UD-Q8_K_XL, Q8_0, … Q4); его тензоры деквантуются, остальные кванты выводятся из них. Если imatrix нет, квантизаторы запускаются с единичными весами или без весов — для каждого тензора выбирается вариант с меньшим остатком. Набор квантов Qwen3.5-2B без BF16 и imatrix (21 файл, 28 ГБ) сжимается до 17%; lmz — до 40%.
+- Если опубликованный imatrix отличается от того, которым квантовали (так в Qwen3.5), часть блоков предсказывается неточно. Корректность от этого не страдает, растёт только размер: для Qwen3.5-2B все кванты вместе стоят 33 МБ.
+- Проверено на Windows 10 и Ubuntu (WSL2) с RTX 5070; другие видеокарты — в плане (ROADMAP, п. 22).
+
+## Лицензия
+
+Apache-2.0 (`LICENSE`). Исходники квантизации llama.cpp / ggml в `seedpack/ggml/` — MIT, см. `NOTICE`.

seedpack-0.1.0/README_PYPI.md

@@ -0,0 +1,70 @@
+<p align="center">
+  <img src="https://raw.githubusercontent.com/glmn-ai/seedpack/main/assets/banner.png" alt="SeedPack" width="860">
+</p>
+
+# SeedPack
+
+**Lossless GPU archiver for GGUF model repositories.** A whole repository, the BF16 model and all its quants from IQ1 to Q8, packs into one archive about the size of a single compressed BF16 file. Every file comes back byte for byte.
+
+| unsloth/Qwen3-0.6B-GGUF, BF16 + 25 quants (10.95 GB) | Archive | Ratio |
+|---|---:|---:|
+| **SeedPack** | **782 MB** | **7.14%** |
+| lmz 1.3.0 | 4.54 GB | 41.47% |
+| zstd 1.5.7 -19 --long=30 | 6.57 GB | 59.99% |
+| 7-Zip 23.01 -mx9 | 10.21 GB | 93.26% |
+| ZipNN 0.5.4 | 10.33 GB | 94.34% |
+
+The archive is smaller than the repository's largest single quant (UD-Q8_K_XL, 844 MB). The full benchmark against 10 other compressors, with speeds and single files, is on the [example archive's page](https://huggingface.co/neurosquad/Qwen3-0.6B-SEEDPACK).
+
+## Install
+
+```bash
+pip install seedpack
+```
+
+Requirements:
+- An NVIDIA GPU, Turing (RTX 20xx) or newer, with a CUDA 13 driver (R580 or newer).
+- Python 3.11+.
+- Windows 10/11 or Linux. Tested on an RTX 5070 under Windows 10 (Python 3.11) and Ubuntu on WSL2 (Python 3.14).
+- No CUDA Toolkit is needed: NVRTC, nvJitLink and the CUDA headers come from pip.
+
+Archives are portable between operating systems: an archive packed on Windows unpacks bit-exactly on Linux, and the other way round. The first run compiles the GPU quantizers for your card (about 15 s) and caches them in `%LOCALAPPDATA%\seedpack\cache` or `~/.cache/seedpack`; set `SEEDPACK_CACHE` to move the cache.
+
+## Use
+
+```bash
+# straight from the Hugging Face Hub: download once, extract what you need
+seedpack list hf://neurosquad/Qwen3-0.6B-SEEDPACK
+seedpack x    hf://neurosquad/Qwen3-0.6B-SEEDPACK Qwen3-0.6B-Q4_K_M.gguf -o models
+seedpack x    hf://neurosquad/Qwen3-0.6B-SEEDPACK "*Q4_K*" "*Q8_0*" -o models
+seedpack x    hf://neurosquad/Qwen3-0.6B-SEEDPACK -o models      # everything
+
+# local archives
+seedpack pack  models/Qwen3-0.6B-GGUF -o qwen3.varc               # files and/or folders
+seedpack x     qwen3.varc Qwen3-0.6B-Q4_K_M.gguf -o out
+seedpack test  qwen3.varc                                         # decode and verify, write nothing
+```
+
+Notes on the commands:
+- `hf://OWNER/REPO[/FILE.varc][@REVISION]` and `https://huggingface.co/...` links work wherever an archive is expected.
+- `list` reads only the archive's index over the network; `x` and `test` download the archive once into the Hugging Face cache.
+- Extracted files are checked against BLAKE3 hashes. Every chunk of the archive carries a CRC32.
+
+## How it works
+
+A GGUF repository ships the same weights 20–30 times: the BF16 model plus each quantization of it. To a general-purpose compressor that looks like gigabytes of almost incompressible bytes. SeedPack uses the fact that every quant is a deterministic function of the same weights:
+
+- **The seed is stored once.** The most precise model (the BF16 file, or the senior quant when there is none) is coded with an entropy coder built for its weights.
+- **Quants are re-derived, not stored.** Each quant is computed again on the GPU by a bit-exact CUDA port of llama.cpp's quantizers (`ggml-quants.c`, compiled without FMA contraction), with the repository's importance matrix.
+- **Only the residual is kept.** Where a re-derived quant differs from the published file, the residual is coded with the quantizer's own probabilities: scales first, then each code relative to its rounding boundary. For Qwen3-0.6B, all 25 quants together cost about 1 MB on top of the seed.
+- **Duplicates are stored once.** Identical tensors in different files are kept once. Files that are not GGUF are compressed with LZMA or zstd.
+
+## Links
+
+- Source and research notes: <https://github.com/glmn-ai/seedpack>
+- Example archive: <https://huggingface.co/neurosquad/Qwen3-0.6B-SEEDPACK>
+- A [NeuroSquad](https://huggingface.co/neurosquad) project.
+
+## License
+
+Apache-2.0. The bundled llama.cpp / ggml quantization sources (`seedpack/ggml/`) are MIT-licensed; see `NOTICE`.

seedpack-0.1.0/pyproject.toml

@@ -0,0 +1,60 @@
+[build-system]
+requires = ["setuptools>=77"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "seedpack"
+dynamic = ["version"]
+description = "Lossless GPU archiver for GGUF model repositories: the BF16 model and all its quants in one archive about the size of a single compressed BF16 file"
+readme = "README_PYPI.md"
+requires-python = ">=3.11"
+license = "Apache-2.0"
+license-files = ["LICENSE", "NOTICE", "seedpack/ggml/LICENSE"]
+authors = [{ name = "NeuroSquad" }]
+keywords = ["gguf", "llama.cpp", "lossless", "compression", "archiver", "quantization", "cuda", "llm", "huggingface"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Environment :: Console",
+    "Environment :: GPU :: NVIDIA CUDA :: 13",
+    "Intended Audience :: Developers",
+    "Intended Audience :: Science/Research",
+    "Operating System :: Microsoft :: Windows",
+    "Operating System :: POSIX :: Linux",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Programming Language :: Python :: 3.14",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
+    "Topic :: System :: Archiving :: Compression",
+]
+dependencies = [
+    "numpy>=2.0",
+    "numba>=0.60",
+    "zstandard>=0.23",
+    "blake3>=1.0",
+    "cupy-cuda13x>=14.0",
+    "nvidia-cuda-nvrtc>=13.0",
+    "nvidia-cuda-runtime>=13.0",
+    "nvidia-nvjitlink>=13.0",
+    "huggingface_hub>=0.30",
+]
+
+[project.urls]
+Homepage = "https://github.com/glmn-ai/seedpack"
+Source = "https://github.com/glmn-ai/seedpack"
+Issues = "https://github.com/glmn-ai/seedpack/issues"
+"Example archive" = "https://huggingface.co/neurosquad/Qwen3-0.6B-SEEDPACK"
+
+[project.scripts]
+seedpack = "seedpack.cli:main"
+
+[tool.setuptools]
+packages = ["seedpack", "seedpack.ggml", "seedpack.ggml.variants"]
+
+[tool.setuptools.dynamic]
+version = { attr = "seedpack.__version__" }
+
+[tool.setuptools.package-data]
+"seedpack.ggml" = ["LICENSE", "*.c", "*.h", "ggml-cpu/*.h", "build.sh", "iqtables.npz"]
+"seedpack.ggml.variants" = ["*.c"]

seedpack-0.1.0/seedpack/__init__.py

@@ -0,0 +1,6 @@
+"""SeedPack: lossless GPU archiver for GGUF model repositories.
+
+The quants of a repository are re-derived on the GPU from its most precise model (the seed) with
+a bit-exact port of llama.cpp's quantisers; only what differs is stored. Command line: `seedpack`.
+"""
+__version__ = "0.1.0"

seedpack-0.1.0/seedpack/__main__.py

@@ -0,0 +1,3 @@
+from .cli import main
+
+main()