Skip to content

tensorflow

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

tensorflow

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
Dockerfile
Dockerfile
 
 
README.md
README.md
 
 
docker-bake.hcl
docker-bake.hcl
 
 

README.md

TensorFlow

A GPU-accelerated TensorFlow container with JupyterLab, pre-configured for the Yotta platform.

What's Included

  • TensorFlow 2.14 with GPU support
  • JupyterLab — interactive notebook environment
  • CUDA — GPU acceleration via NVIDIA runtime
  • System toolchain: SSH, Nginx, git, wget, curl

Exposed Ports

Port Protocol Service
22 TCP SSH
8888 HTTP JupyterLab

Environment Variables

Variable Default Description
JUPYTER_PASSWORD ubuntu JupyterLab login password

Quick Start

Verify TensorFlow and GPU

import tensorflow as tf

print(tf.__version__)
print("GPUs available:", tf.config.list_physical_devices('GPU'))

Build Instructions

From the repository root:

# Build and push
docker buildx bake tensorflow --push

# Build without cache
docker buildx bake tensorflow --no-cache --push

Notes

  • Mount /workspace as a persistent volume to retain notebooks and data across restarts.
  • Change the default JupyterLab password by setting the JUPYTER_PASSWORD environment variable.