ggufkit 0.1.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- ggufkit-0.1.0/.gitignore +171 -0
- ggufkit-0.1.0/CHANGELOG.md +56 -0
- ggufkit-0.1.0/CITATION.cff +23 -0
- ggufkit-0.1.0/LICENSE +22 -0
- ggufkit-0.1.0/PKG-INFO +417 -0
- ggufkit-0.1.0/README.md +383 -0
- ggufkit-0.1.0/SECURITY.md +37 -0
- ggufkit-0.1.0/docs/FINGERPRINT.md +168 -0
- ggufkit-0.1.0/docs/MANIFEST.md +159 -0
- ggufkit-0.1.0/docs/RELEASING.md +75 -0
- ggufkit-0.1.0/pyproject.toml +140 -0
- ggufkit-0.1.0/src/ggufkit/__init__.py +12 -0
- ggufkit-0.1.0/src/ggufkit/__main__.py +5 -0
- ggufkit-0.1.0/src/ggufkit/_util.py +75 -0
- ggufkit-0.1.0/src/ggufkit/canon.py +200 -0
- ggufkit-0.1.0/src/ggufkit/cli.py +80 -0
- ggufkit-0.1.0/src/ggufkit/dequant.py +70 -0
- ggufkit-0.1.0/src/ggufkit/diff.py +462 -0
- ggufkit-0.1.0/src/ggufkit/fingerprint.py +263 -0
- ggufkit-0.1.0/src/ggufkit/manifest.py +556 -0
- ggufkit-0.1.0/src/ggufkit/model.py +301 -0
- ggufkit-0.1.0/src/ggufkit/py.typed +0 -0
- ggufkit-0.1.0/src/ggufkit/stitch.py +163 -0
- ggufkit-0.1.0/src/ggufkit/verify.py +426 -0
- ggufkit-0.1.0/src/ggufkit/viz/__init__.py +28 -0
- ggufkit-0.1.0/src/ggufkit/viz/errmap.py +374 -0
- ggufkit-0.1.0/src/ggufkit/viz/layers.py +410 -0
- ggufkit-0.1.0/src/ggufkit/viz/palette.py +71 -0
- ggufkit-0.1.0/src/ggufkit/viz/render.py +408 -0
- ggufkit-0.1.0/tests/__init__.py +0 -0
- ggufkit-0.1.0/tests/conftest.py +187 -0
- ggufkit-0.1.0/tests/golden/tiny.json +59 -0
- ggufkit-0.1.0/tests/golden/weights.json +7 -0
- ggufkit-0.1.0/tests/test_canon.py +90 -0
- ggufkit-0.1.0/tests/test_cli.py +83 -0
- ggufkit-0.1.0/tests/test_diff.py +158 -0
- ggufkit-0.1.0/tests/test_fingerprint.py +175 -0
- ggufkit-0.1.0/tests/test_golden.py +73 -0
- ggufkit-0.1.0/tests/test_manifest.py +248 -0
- ggufkit-0.1.0/tests/test_model.py +52 -0
- ggufkit-0.1.0/tests/test_robustness.py +302 -0
- ggufkit-0.1.0/tests/test_stitch.py +54 -0
- ggufkit-0.1.0/tests/test_verify.py +238 -0
- ggufkit-0.1.0/tests/test_viz.py +171 -0
- ggufkit-0.1.0/tests/test_viz_error.py +188 -0
- ggufkit-0.1.0/tests/test_viz_layers.py +191 -0
ggufkit-0.1.0/.gitignore
ADDED
|
@@ -0,0 +1,171 @@
|
|
|
1
|
+
# Byte-compiled / optimized / DLL files
|
|
2
|
+
__pycache__/
|
|
3
|
+
*.py[cod]
|
|
4
|
+
*$py.class
|
|
5
|
+
|
|
6
|
+
# C extensions
|
|
7
|
+
*.so
|
|
8
|
+
|
|
9
|
+
# Distribution / packaging
|
|
10
|
+
.Python
|
|
11
|
+
build/
|
|
12
|
+
develop-eggs/
|
|
13
|
+
dist/
|
|
14
|
+
downloads/
|
|
15
|
+
eggs/
|
|
16
|
+
.eggs/
|
|
17
|
+
lib/
|
|
18
|
+
lib64/
|
|
19
|
+
parts/
|
|
20
|
+
sdist/
|
|
21
|
+
var/
|
|
22
|
+
wheels/
|
|
23
|
+
share/python-wheels/
|
|
24
|
+
*.egg-info/
|
|
25
|
+
.installed.cfg
|
|
26
|
+
*.egg
|
|
27
|
+
MANIFEST
|
|
28
|
+
|
|
29
|
+
# PyInstaller
|
|
30
|
+
# Usually these files are written by a python script from a template
|
|
31
|
+
# before PyInstaller builds the exe, so as to inject date/other infos into it.
|
|
32
|
+
*.manifest
|
|
33
|
+
*.spec
|
|
34
|
+
|
|
35
|
+
# Installer logs
|
|
36
|
+
pip-log.txt
|
|
37
|
+
pip-delete-this-directory.txt
|
|
38
|
+
|
|
39
|
+
# Unit test / coverage reports
|
|
40
|
+
htmlcov/
|
|
41
|
+
.tox/
|
|
42
|
+
.nox/
|
|
43
|
+
.coverage
|
|
44
|
+
.coverage.*
|
|
45
|
+
.cache
|
|
46
|
+
nosetests.xml
|
|
47
|
+
coverage.xml
|
|
48
|
+
*.cover
|
|
49
|
+
*.py,cover
|
|
50
|
+
.hypothesis/
|
|
51
|
+
.pytest_cache/
|
|
52
|
+
cover/
|
|
53
|
+
|
|
54
|
+
# Translations
|
|
55
|
+
*.mo
|
|
56
|
+
*.pot
|
|
57
|
+
|
|
58
|
+
# Django stuff:
|
|
59
|
+
*.log
|
|
60
|
+
local_settings.py
|
|
61
|
+
db.sqlite3
|
|
62
|
+
db.sqlite3-journal
|
|
63
|
+
|
|
64
|
+
# Flask stuff:
|
|
65
|
+
instance/
|
|
66
|
+
.webassets-cache
|
|
67
|
+
|
|
68
|
+
# Scrapy stuff:
|
|
69
|
+
.scrapy
|
|
70
|
+
|
|
71
|
+
# Sphinx documentation
|
|
72
|
+
docs/_build/
|
|
73
|
+
|
|
74
|
+
# PyBuilder
|
|
75
|
+
.pybuilder/
|
|
76
|
+
target/
|
|
77
|
+
|
|
78
|
+
# Jupyter Notebook
|
|
79
|
+
.ipynb_checkpoints
|
|
80
|
+
|
|
81
|
+
# IPython
|
|
82
|
+
profile_default/
|
|
83
|
+
ipython_config.py
|
|
84
|
+
|
|
85
|
+
# pyenv
|
|
86
|
+
# For a library or package, you might want to ignore these files since the code is
|
|
87
|
+
# intended to run in multiple environments; otherwise, check them in:
|
|
88
|
+
# .python-version
|
|
89
|
+
|
|
90
|
+
# pipenv
|
|
91
|
+
# According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
|
|
92
|
+
# However, in case of collaboration, if having platform-specific dependencies or dependencies
|
|
93
|
+
# having no cross-platform support, pipenv may install dependencies that don't work, or not
|
|
94
|
+
# install all needed dependencies.
|
|
95
|
+
#Pipfile.lock
|
|
96
|
+
|
|
97
|
+
# poetry
|
|
98
|
+
# Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
|
|
99
|
+
# This is especially recommended for binary packages to ensure reproducibility, and is more
|
|
100
|
+
# commonly ignored for libraries.
|
|
101
|
+
# https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
|
|
102
|
+
#poetry.lock
|
|
103
|
+
|
|
104
|
+
# pdm
|
|
105
|
+
# Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
|
|
106
|
+
#pdm.lock
|
|
107
|
+
# pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
|
|
108
|
+
# in version control.
|
|
109
|
+
# https://pdm.fming.dev/#use-with-ide
|
|
110
|
+
.pdm.toml
|
|
111
|
+
|
|
112
|
+
# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
|
|
113
|
+
__pypackages__/
|
|
114
|
+
|
|
115
|
+
# Celery stuff
|
|
116
|
+
celerybeat-schedule
|
|
117
|
+
celerybeat.pid
|
|
118
|
+
|
|
119
|
+
# SageMath parsed files
|
|
120
|
+
*.sage.py
|
|
121
|
+
|
|
122
|
+
# Environments
|
|
123
|
+
.env
|
|
124
|
+
.venv
|
|
125
|
+
env/
|
|
126
|
+
venv/
|
|
127
|
+
ENV/
|
|
128
|
+
env.bak/
|
|
129
|
+
venv.bak/
|
|
130
|
+
|
|
131
|
+
# Spyder project settings
|
|
132
|
+
.spyderproject
|
|
133
|
+
.spyproject
|
|
134
|
+
|
|
135
|
+
# Rope project settings
|
|
136
|
+
.ropeproject
|
|
137
|
+
|
|
138
|
+
# mkdocs documentation
|
|
139
|
+
/site
|
|
140
|
+
|
|
141
|
+
# mypy
|
|
142
|
+
.mypy_cache/
|
|
143
|
+
.dmypy.json
|
|
144
|
+
dmypy.json
|
|
145
|
+
|
|
146
|
+
# Pyre type checker
|
|
147
|
+
.pyre/
|
|
148
|
+
|
|
149
|
+
# pytype static type analyzer
|
|
150
|
+
.pytype/
|
|
151
|
+
|
|
152
|
+
# Cython debug symbols
|
|
153
|
+
cython_debug/
|
|
154
|
+
|
|
155
|
+
# PyCharm
|
|
156
|
+
# JetBrains specific template is maintained in a separate JetBrains.gitignore that can
|
|
157
|
+
# be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
|
|
158
|
+
# and can be added to the global gitignore or merged into this file. For a more nuclear
|
|
159
|
+
# option (not recommended) you can uncomment the following to ignore the entire idea folder.
|
|
160
|
+
#.idea/
|
|
161
|
+
|
|
162
|
+
# ggufkit
|
|
163
|
+
*.gguf
|
|
164
|
+
*.manifest.json
|
|
165
|
+
/scratch/
|
|
166
|
+
.ruff_cache/
|
|
167
|
+
# Local signing runs (scripts/make_demos.py makes throwaway keys): never commit these.
|
|
168
|
+
*.key
|
|
169
|
+
*.sigstore.json
|
|
170
|
+
*.intoto.json
|
|
171
|
+
node_modules/
|
|
@@ -0,0 +1,56 @@
|
|
|
1
|
+
# Changelog
|
|
2
|
+
|
|
3
|
+
All notable changes to this project are documented here. The format follows
|
|
4
|
+
[Keep a Changelog](https://keepachangelog.com/en/1.1.0/) and the project uses
|
|
5
|
+
[semantic versioning](https://semver.org/).
|
|
6
|
+
|
|
7
|
+
## [0.1.0] - 2026-10-27
|
|
8
|
+
|
|
9
|
+
First release. Supports Python 3.11 to 3.14.
|
|
10
|
+
|
|
11
|
+
### Added
|
|
12
|
+
|
|
13
|
+
- `ggufkit fingerprint`: order-independent content fingerprint (`ggufkit-fp/1`, see
|
|
14
|
+
`docs/FINGERPRINT.md`) with separate digests for metadata, tensor descriptors and weights.
|
|
15
|
+
Invariant to metadata and tensor order, offsets, alignment, sharding and, for element types,
|
|
16
|
+
byte order. `weights` binds each tensor's bytes to its name, so data moved between tensors
|
|
17
|
+
is a change.
|
|
18
|
+
- `ggufkit manifest`: provenance record (`ggufkit-manifest/1`, see `docs/MANIFEST.md`) with the
|
|
19
|
+
fingerprint, file hashes, metadata claims, user-supplied build inputs and a provenance gap list.
|
|
20
|
+
- `ggufkit verify`: compare a model with a manifest or another file and get a one-word verdict
|
|
21
|
+
with a meaningful exit status. It passes only when the weights were actually compared (or
|
|
22
|
+
`--no-data` asks for a metadata-only check), and `--file-sha256` passes only when the
|
|
23
|
+
manifest recorded file hashes.
|
|
24
|
+
- `ggufkit diff`: metadata value diff, tensor diff and `--numeric` per-tensor error statistics.
|
|
25
|
+
- `ggufkit stitch`: combine the metadata of one file with the tensors of another, or write a
|
|
26
|
+
vocab-only file.
|
|
27
|
+
- `ggufkit viz tensor`: render tensors of any type `gguf.quants` can decode as images.
|
|
28
|
+
- Split models (`-00001-of-00003.gguf`) are handled as one model everywhere.
|
|
29
|
+
- `ggufkit viz error`: per-weight error heatmaps between two models' tensors (absolute,
|
|
30
|
+
relative or signed), pooled to a pixel budget while decoding, with a caption of rmse,
|
|
31
|
+
relative L2 and cosine similarity, and an optional shared colour scale.
|
|
32
|
+
- `ggufkit viz layers`: a grid of the quantization type of every tensor, block by block,
|
|
33
|
+
coloured by bits per weight; also as a text table or JSON.
|
|
34
|
+
- `ggufkit viz layers` takes several models and draws them side by side with one palette,
|
|
35
|
+
so a colour means the same type in every panel.
|
|
36
|
+
- README images and examples from a real model (Qwen3-0.6B, BF16, Q4_K_M and UD-Q4_K_XL from
|
|
37
|
+
unsloth/Qwen3-0.6B-GGUF), regenerated by `scripts/make_readme_assets.py` from a pinned
|
|
38
|
+
revision with sha256 checks. `scripts/make_examples.py` keeps one synthetic illustration.
|
|
39
|
+
- `ggufkit manifest --intoto FILE`: an in-toto Statement v1 for signing; `docs/MANIFEST.md`
|
|
40
|
+
has a cosign recipe tested on a real model.
|
|
41
|
+
- Malformed and hostile files fail fast with one `error:` line and exit status 2: a header
|
|
42
|
+
check rejects declared sizes that cannot fit in the file before gguf-py parses it. Output
|
|
43
|
+
files are written atomically, and existing ones are only replaced with `--force`.
|
|
44
|
+
- `python -m ggufkit`, and type information for the package (`py.typed`).
|
|
45
|
+
- README demo animations (`scripts/make_demos.py`, `scripts/demos/`), recorded from real
|
|
46
|
+
commands on the same files, with 1080x1920 and 4:5 MP4 renders and SRT captions for posting
|
|
47
|
+
elsewhere.
|
|
48
|
+
- Releases are built once, smoke-tested from a clean install, attested (build provenance and
|
|
49
|
+
PyPI attestations) and published by `.github/workflows/release.yml`.
|
|
50
|
+
|
|
51
|
+
### Origins
|
|
52
|
+
|
|
53
|
+
Builds on [KerfuffleV2/gguf-tools](https://github.com/KerfuffleV2/gguf-tools) (MIT, 2023):
|
|
54
|
+
`gguf-checksum`, `gguf-frankenstein` and `gguf-tensor-to-image`, whose history is preserved.
|
|
55
|
+
|
|
56
|
+
[0.1.0]: https://github.com/MysterionRise/ggufkit/releases/tag/v0.1.0
|
|
@@ -0,0 +1,23 @@
|
|
|
1
|
+
cff-version: 1.2.0
|
|
2
|
+
message: "If you use ggufkit in your work, please cite it as below."
|
|
3
|
+
title: "ggufkit: fingerprint, verify and visualize GGUF models"
|
|
4
|
+
type: software
|
|
5
|
+
authors:
|
|
6
|
+
- family-names: Perikov
|
|
7
|
+
given-names: Konstantin
|
|
8
|
+
- alias: KerfuffleV2
|
|
9
|
+
name: Kerfuffle
|
|
10
|
+
version: 0.1.0
|
|
11
|
+
date-released: "2026-10-27"
|
|
12
|
+
license: MIT
|
|
13
|
+
repository-code: "https://github.com/MysterionRise/ggufkit"
|
|
14
|
+
keywords:
|
|
15
|
+
- gguf
|
|
16
|
+
- llama.cpp
|
|
17
|
+
- quantization
|
|
18
|
+
- provenance
|
|
19
|
+
- model signing
|
|
20
|
+
abstract: >-
|
|
21
|
+
ggufkit gives GGUF model files an order-independent content fingerprint, a provenance
|
|
22
|
+
manifest that records what a file claims and what it leaves out, verification against either,
|
|
23
|
+
and images of quantization error and quantization recipes.
|
ggufkit-0.1.0/LICENSE
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
1
|
+
MIT License
|
|
2
|
+
|
|
3
|
+
Copyright (c) 2023 Kerfuffle (original gguf-tools scripts)
|
|
4
|
+
Copyright (c) 2026 Konstantin Perikov (ggufkit)
|
|
5
|
+
|
|
6
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
7
|
+
of this software and associated documentation files (the "Software"), to deal
|
|
8
|
+
in the Software without restriction, including without limitation the rights
|
|
9
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
10
|
+
copies of the Software, and to permit persons to whom the Software is
|
|
11
|
+
furnished to do so, subject to the following conditions:
|
|
12
|
+
|
|
13
|
+
The above copyright notice and this permission notice shall be included in all
|
|
14
|
+
copies or substantial portions of the Software.
|
|
15
|
+
|
|
16
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
17
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
18
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
19
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
20
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
21
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
22
|
+
SOFTWARE.
|