Julia 1.0.1 Notebook gets crash


#1

I have issue with the below code and my jupyter notebook keep getting crash. Any help is really appreciate.

using Pkg
Pkg.add("Flux")
Pkg.add("ImageMagick")
Pkg.add("CSV")
Pkg.add("DataFrames")
Pkg.add("CSVFiles")
using Flux, Flux.Data.MNIST
using Flux: onehotbatch, argmax, crossentropy, throttle
using Base.Iterators: repeated
using Images
using CSV
using DataFrames
using CSVFiles

data_train = load("train.csv") |> DataFrame

train.csv is MNIST training data from Kaggle which is of size (60000,765).


#2

Try CSV.read(“train.csv”)


#3

Thanks for your response.

I have tried this but the performance is very slow with this approach. It takes almost more than 30 mins to load 42000 X 765 data.


#4

Hi Rama,
Welcome to Discourse! For me, running

data_train = load("train.csv") |> DataFrame

with that train.csv file from Kaggle works on julia 1.0.1. I didn’t try it from a notebook but I can’t imagine that’s the issue. It’s hard to say what the problem might be without more information. Could you post a bit more information on the crash? Are you seeing a Julia error message, in which case, could you post the backtrace? Or is your notebook just dying on you without a useful error message?

Also, for me CSV.read(“train.csv”) takes just a couple of seconds. The fact that it’s taking more than 30 minutes for you would seem to be either an indication that something’s wrong with your julia or CSV.jl installation or that you weren’t doing what you thought you were doing.

Cheers, Patrick


#5

Thanks for your response.

My notebook just dying without a useful error message.

so I changed my approach and used Pandas and it was very quick loading the same file in pandas with the below code.

data_train = Pandas.read_csv(“train.csv”)

Thanks again.

Thanks
Rama