how to install h2o4gpu on colab and use colab gpu for Machine Learning algorithm

Install h2o4gpu and tpot on google colab (GPU)

# How To Install h2o4gpu and Tpot in colab?
1. export some env variable
2. install linux packges
3. uninstall sklearn and install python packges
4. enjoy fast auto ML with gpu

install-h2o4gpu-colab.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# How To Install h2o4gpu and Tpot in colab?\n",
    "1. export some env variable\n",
    "2. install linux packges\n",
    "3. uninstall sklearn and install python packges\n",
    "4. enjoy fast auto ML with gpu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# export some env variable\n",
    "!export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$CUDA_HOME/lib64/:$CUDA_HOME/lib/:$CUDA_HOME/extras/CUPTI/lib64"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#install python packges\n",
    "!sudo apt-get install libopenblas-dev pbzip2\n",
    "!sudo apt-get -y install libcurl4-openssl-dev libssl-dev libxml2-dev"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# uninstall sklearn and install python packges\n",
    "!pip uninstall sklearn\n",
    "!pip install -U --force-reinstall ipykernel\n",
    "!pip install parallel\n",
    "!pip install --force-reinstall -I h2o4gpu\n",
    "!pip install TPOT"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## not need to reset your runtime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# enjoy it\n",
    "from tpot import TPOTClassifier\n",
    "from h2o4gpu.datasets import load_iris\n",
    "from h2o4gpu.model_selection import train_test_split\n",
    "import numpy as np\n",
    "\n",
    "iris = load_iris()\n",
    "X_train, X_test, y_train, y_test = train_test_split(iris.data.astype(np.float64),\n",
    "    iris.target.astype(np.float64), train_size=0.75, test_size=0.25, random_state=42)\n",
    "\n",
    "tpot = TPOTClassifier(generations=5, population_size=50, verbosity=2, random_state=42)\n",
    "tpot.fit(X_train, y_train)\n",
    "print(tpot.score(X_test, y_test))\n",
    "tpot.export('tpot_iris_pipeline.py')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "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
}

I tried this solution but unfortunately it didn't work. Is it still viable?