Docker Image for TensorFlow & PyTorch GPU
Budget: $10 – $30 USD
I am looking for a professional who can create a Docker image tailored for machine learning development. This image should come equipped with Python 3.10.8, the latest stable releases of TensorFlow and PyTorch, and should utilize an environment.yml for managing dependencies.
Objective: Create a Docker image that installs all necessary components to run Python scripts with TensorFlow (GPU version) and PyTorch (GPU version), along with handling dependencies specified in an `environment.yml` file.
- Deliverables
1. Dockerfile: A complete Dockerfile that:
- Uses an appropriate base image (e.g., `nvidia/cuda`) with CUDA and cuDNN installed.
- Installs Python (preferably 3.10.8 or compatible).
- Installs TensorFlow with GPU support and PyTorch with GPU support.
- Installs dependencies from an `environment.yml` file.
- Sets up a working directory for the application.
- Copies the necessary files into the container.
- Specifies a default command to run the Python script.
2. Automation Script: A shell script (e.g., `build_and_run.sh`) that:
- Builds the Docker image.
- Runs the Docker container with GPU support, ensuring proper volume mapping for your project files.
- Technical Specifications
- Base Image: Use `nvidia/cuda:11.8.0-cudnn8-runtime-ubuntu20.04` (or latest compatible version).
- Python Version: Python 3.10.8.
- TensorFlow Version: Use the latest compatible version with GPU support.
- PyTorch Version: Use the latest compatible version with GPU support.
- Dependencies: Install packages listed in `environment.yml`.
- Additional Requirements
- Ensure the Dockerfile is well-documented with comments explaining each step.
- Test the Docker image to confirm that:
- Python scripts run without errors.
- TensorFlow and PyTorch can utilize the GPU.
- Ensure the automation script includes:
- Building the Docker image with an appropriate tag.
- Running the Docker container with necessary options (e.g., --gpus all).
- Volume mapping to allow access to project files from the host.
- Option to specify environment variables if needed.
- Communication
- Regular updates on progress and any challenges encountered.
- Notify upon completion of the project and provide documentation.
Ideal skills and experience for this job include:
- Proficiency in Docker and Python
- Experience with TensorFlow and PyTorch
- Ability to create automated scripts
- Strong documentation and testing skills
Please specify your budget and timeline in your proposal.
Objective: Create a Docker image that installs all necessary components to run Python scripts with TensorFlow (GPU version) and PyTorch (GPU version), along with handling dependencies specified in an `environment.yml` file.
- Deliverables
1. Dockerfile: A complete Dockerfile that:
- Uses an appropriate base image (e.g., `nvidia/cuda`) with CUDA and cuDNN installed.
- Installs Python (preferably 3.10.8 or compatible).
- Installs TensorFlow with GPU support and PyTorch with GPU support.
- Installs dependencies from an `environment.yml` file.
- Sets up a working directory for the application.
- Copies the necessary files into the container.
- Specifies a default command to run the Python script.
2. Automation Script: A shell script (e.g., `build_and_run.sh`) that:
- Builds the Docker image.
- Runs the Docker container with GPU support, ensuring proper volume mapping for your project files.
- Technical Specifications
- Base Image: Use `nvidia/cuda:11.8.0-cudnn8-runtime-ubuntu20.04` (or latest compatible version).
- Python Version: Python 3.10.8.
- TensorFlow Version: Use the latest compatible version with GPU support.
- PyTorch Version: Use the latest compatible version with GPU support.
- Dependencies: Install packages listed in `environment.yml`.
- Additional Requirements
- Ensure the Dockerfile is well-documented with comments explaining each step.
- Test the Docker image to confirm that:
- Python scripts run without errors.
- TensorFlow and PyTorch can utilize the GPU.
- Ensure the automation script includes:
- Building the Docker image with an appropriate tag.
- Running the Docker container with necessary options (e.g., --gpus all).
- Volume mapping to allow access to project files from the host.
- Option to specify environment variables if needed.
- Communication
- Regular updates on progress and any challenges encountered.
- Notify upon completion of the project and provide documentation.
Ideal skills and experience for this job include:
- Proficiency in Docker and Python
- Experience with TensorFlow and PyTorch
- Ability to create automated scripts
- Strong documentation and testing skills
Please specify your budget and timeline in your proposal.