Skip to content

Installation

The MLdebugger Python SDK is a framework for collecting internal features and metadata from models built and trained in Python, enabling evaluation and debugging.

System Requirements

  • Python: 3.7 - 3.12
  • ML Framework: PyTorch (1.7.1-2.10.0) or TensorFlow/Keras (2.1.0-2.18.0)

About Python 3.6

Python 3.6 is not supported. Core dependencies (cryptography, SQLAlchemy 2.x, pandas 1.3+) require Python 3.7+.

Prerequisites

This SDK assumes that inference can be executed using models implemented in PyTorch or TensorFlow. Please run the following commands in a Python environment where one of these ML frameworks is installed and the model can be executed.

Installation Methods

Standard Installation

Install directly using the Dropbox shared link.

pip install "https://dl.dropboxusercontent.com/s/hw78mfecqq3qyy8630d15/ml_debugger-0.3.3-py3-none-any.whl?rlkey=00o6520zstd9syaosw2rxypz9"

Installation URL

The installation URL may change. Contact the Adansons team for the latest URL.

Installation Without Dependencies

If you want to avoid dependency conflicts with your existing environment, use the --no-deps option.

pip install --no-deps "https://dl.dropboxusercontent.com/s/hw78mfecqq3qyy8630d15/ml_debugger-0.3.3-py3-none-any.whl?rlkey=00o6520zstd9syaosw2rxypz9"

In this case, you need to install the following main dependencies separately:

Package Minimum Version Purpose
numpy 1.19.0 Numerical computation
pandas 1.3.0 Data manipulation
sqlalchemy 1.4.0 Database operations
sqlmodel 0.0.8 ORM models
requests 2.25.0 API communication
pyyaml 6.0 Configuration files

Update Method

To update an existing installation to the latest version, use the --upgrade option.

pip install --upgrade "https://dl.dropboxusercontent.com/s/hw78mfecqq3qyy8630d15/ml_debugger-0.3.3-py3-none-any.whl?rlkey=00o6520zstd9syaosw2rxypz9"

Installation Verification

To verify that the installation was successful, run the following command.

python -c "from importlib.metadata import version; print(version('ml-debugger'))"

If it runs without errors and displays the version number, the installation is successful.

0.3.3

Next Steps

After installation is complete, proceed to Getting Started to learn the basic usage of the SDK.