From the ML forum, Colab instance crashing due to RAM

vishak · December 8, 2019, 1:07pm

In Lesson 3 of ML, Jeremy is able to apply

add_datepart(df_all,‘date’)

on the grocery dataset. I tried this and the runtime crashed. It was a 12GB RAM machine on colab.
Switching to a larger 25GB Ram didn’t help. The session crashed.

What exactly is happening? Please help!

Here’s the link to that notebook on github
The last cell is the one that causing the crash

github.com

VishakBharadwaj94/Corporaci-n_Favorita_Grocery_Sales_Forecasting/blob/master/Corporación_Favorita_Grocery_Sales_Forecasting.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Corporación_Favorita_Grocery_Sales_Forecasting",
      "provenance": [],
      "machine_shape": "hm",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",

This file has been truncated. show original

chatuur · December 8, 2019, 1:30pm

Try reducing the num_workers to 0 or 1 when creating a databunch.

vishak · December 8, 2019, 3:03pm

Hey,the issue is a traditional ML issue. I’m not using the fastai databunch for this. I’m using pandas to work on the data and then fit to a random forest algorithm