Build a machine learning development environment, with WSL2 using DirectML and TensorFlow
According to Microsoft link, to create a machine learning development environment with WSL2 on Windows using DirectML and TensorFlow, we need to follow these steps:
- Install the Latest GPU Driver: Download and install the latest GPU driver for our hardware
- Enable WSL and Install a glibc-based Distribution
- Install Miniconda inside our WSL instance and set up a virtual Python environment using Miniconda
- Create a Conda Environment: Create an environment using Python and activate it
- Install the TensorFlow with DirectML package through pip
Lot of steps right?
No worries, there's a 🍭Candy for that: ub-python-miniconda-tensorflow-directml
This time-saving Candy "package" simplifies the entire process, combining Miniconda, TensorFlow, and DirectML in a convenient package.
We only need to install and start using it.
Add-PnPWsl2Candy -Candy ub-python-Miniconda-TensorFlow-DirectML -Instance mywslinstance
The "package" will install the following components:
Miniconda is a small bootstrap version of Anaconda: includes conda, Python, usefull pkgs
TensorFlow is a popular open-source machine learning framework. Its is a symbolic math library based on dataflow and differentiable programming, used for machine learning applications such as neural networks.
Direct Machine Learning (DirectML) is a low-level API for machine learning (ML). Hardware-accelerated machine learning primitives (called operators) are the building blocks of DirectML. From those building blocks, we can develop such machine learning techniques as upscaling, anti-aliasing, and style transfer, to name but a few.