TensorFlow を Geforce RTX5000 シリーズで動かす

Page content

TensorFlow の 2.20.0 が 8/14 にリリースされたが、 Geforce RTX5000 シリーズは対応されていなかった。

ただ nightly だと動くことが分ったので、 ここでは nightly を使った uv 用の pyproject.toml と TensorFlow を使ったサンプルを載せておく。

Geforce RTX5000 シリーズ対応の TensorFlow をセットアップする uv 用の pyproject.toml

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
[project]
name = "note-base-tf"
version = "0.1.0"
description = "Add your description here"
readme = "README.md"
requires-python = ">=3.11.5"
dependencies = [
    "datasets>=3.0.1",
    "flax>=0.10.6",
    "image-classifiers>=1.0.0",
    "ipywidgets>=8.1.7",
    "keras>=3.10.0",
    "matplotlib>=3.9.0",
    "notebook>=7.4.3",
    "nvidia-cudnn-cu12>=9.12.0.46",
    "opencv-python>=4.11.0.86",
    "pillow>=11.2.1",
    "scikit-learn>=1.6.1",
    "tensorboardx>=2.6.2.2",
    "tf-nightly>=2.21.0.dev20250829",
]