vramfit 0.0.1__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.
- vramfit-0.0.1/PKG-INFO +26 -0
- vramfit-0.0.1/README.md +13 -0
- vramfit-0.0.1/pyproject.toml +25 -0
- vramfit-0.0.1/src/vramfit/__init__.py +9 -0
vramfit-0.0.1/PKG-INFO
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
1
|
+
Metadata-Version: 2.5
|
|
2
|
+
Name: vramfit
|
|
3
|
+
Version: 0.0.1
|
|
4
|
+
Summary: Fit large open models on a single GPU: per-layer quantization sensitivity, mixed-precision recipes under a VRAM budget. Name claim ahead of the first release.
|
|
5
|
+
Project-URL: Repository, https://github.com/Alberto-Codes/quantfit
|
|
6
|
+
Author-email: Alberto-Codes <alberto.codes.dev@gmail.com>
|
|
7
|
+
License: MIT
|
|
8
|
+
Classifier: Development Status :: 1 - Planning
|
|
9
|
+
Classifier: License :: OSI Approved :: MIT License
|
|
10
|
+
Classifier: Programming Language :: Python :: 3 :: Only
|
|
11
|
+
Requires-Python: >=3.12
|
|
12
|
+
Description-Content-Type: text/markdown
|
|
13
|
+
|
|
14
|
+
# vramfit
|
|
15
|
+
|
|
16
|
+
Fit large open models on a single GPU.
|
|
17
|
+
|
|
18
|
+
vramfit measures per-layer quantization sensitivity, solves for a
|
|
19
|
+
mixed-precision recipe under a VRAM budget, validates the recipe
|
|
20
|
+
against the prediction, and packs the result for a target runtime.
|
|
21
|
+
|
|
22
|
+
This 0.0.1 release claims the name ahead of the first functional
|
|
23
|
+
release. The tool is real and public today under its former name:
|
|
24
|
+
<https://github.com/Alberto-Codes/quantfit>. The repository renames
|
|
25
|
+
to vramfit and the first functional release replaces this
|
|
26
|
+
placeholder shortly.
|
vramfit-0.0.1/README.md
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
1
|
+
# vramfit
|
|
2
|
+
|
|
3
|
+
Fit large open models on a single GPU.
|
|
4
|
+
|
|
5
|
+
vramfit measures per-layer quantization sensitivity, solves for a
|
|
6
|
+
mixed-precision recipe under a VRAM budget, validates the recipe
|
|
7
|
+
against the prediction, and packs the result for a target runtime.
|
|
8
|
+
|
|
9
|
+
This 0.0.1 release claims the name ahead of the first functional
|
|
10
|
+
release. The tool is real and public today under its former name:
|
|
11
|
+
<https://github.com/Alberto-Codes/quantfit>. The repository renames
|
|
12
|
+
to vramfit and the first functional release replaces this
|
|
13
|
+
placeholder shortly.
|
|
@@ -0,0 +1,25 @@
|
|
|
1
|
+
[project]
|
|
2
|
+
name = "vramfit"
|
|
3
|
+
version = "0.0.1"
|
|
4
|
+
description = "Fit large open models on a single GPU: per-layer quantization sensitivity, mixed-precision recipes under a VRAM budget. Name claim ahead of the first release."
|
|
5
|
+
readme = "README.md"
|
|
6
|
+
license = { text = "MIT" }
|
|
7
|
+
authors = [
|
|
8
|
+
{ name = "Alberto-Codes", email = "alberto.codes.dev@gmail.com" }
|
|
9
|
+
]
|
|
10
|
+
requires-python = ">=3.12"
|
|
11
|
+
classifiers = [
|
|
12
|
+
"Development Status :: 1 - Planning",
|
|
13
|
+
"License :: OSI Approved :: MIT License",
|
|
14
|
+
"Programming Language :: Python :: 3 :: Only",
|
|
15
|
+
]
|
|
16
|
+
|
|
17
|
+
[project.urls]
|
|
18
|
+
Repository = "https://github.com/Alberto-Codes/quantfit"
|
|
19
|
+
|
|
20
|
+
[build-system]
|
|
21
|
+
requires = ["hatchling"]
|
|
22
|
+
build-backend = "hatchling.build"
|
|
23
|
+
|
|
24
|
+
[tool.hatch.build.targets.wheel]
|
|
25
|
+
packages = ["src/vramfit"]
|