Azure Machine Learning environment as Conda kernel in Jupyter / Jupyterlab

AML Environment as Conda kernel.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create an AzureML Environment as a local Conda kernel in your Jupyter / JupyterLab"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1. Download the Conda Dependencies file from AzureML"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "from azureml.core import Workspace, Environment\n",
    "\n",
    "ws = Workspace.from_config()\n",
    "\n",
    "tf_env = Environment.get(ws, 'AzureML-TensorFlow-2.0-GPU')\n",
    "tf_env.python.conda_dependencies.save(\"./CondaDependencies.yaml\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "CONDA_ENV_NAME = 'amltf2'\n",
    "KERNEL_DISPLAY_NAME = 'AzureML - TensorFlow-2.0 - GPU'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2. Create the Conda Environment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "%conda env create --name {CONDA_ENV_NAME} --file ./CondaDependencies.yaml"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3. Install ipykernel in the new Conda environment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "%conda install -n {CONDA_ENV_NAME} ipykernel -y"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4. Run ipykernel in the new Conda environment to register the kernelspec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!/anaconda/envs/{CONDA_ENV_NAME}/bin/python -m ipykernel install --user --name {CONDA_ENV_NAME} --display-name \"{KERNEL_DISPLAY_NAME}\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5. Wait for a few minutes and the new kernel will show up in Jupyter/Jupyterlab under the available kernels"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.6 - AzureML",
   "language": "python",
   "name": "python3-azureml"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}

Hello @ezwiefel ,

I've tried to setup AzureML Env as Conda Kernel in Jupiter but unfornunately i have an error message on jupiter notebook user interface on AML depiste being able to correctly activate env on terminal interface

FYI, my Azure ML workspace is behind a VNET