Lesson 1 Notebook - can't find dogscats data

(Simon Laycock) #1

Hi - please bear with a complete newbie! I have managed to get through various obstacles myself, but now I’m stuck. I wonder if I have unzipped the dogscats data into the wrong directory, but it looks to be exactly the same as in the video. It doesn’t help that I know very little of Linux. Anyhow, the notebook runs fine until the final step, when I get:

OSError: [Errno 2] No such file or directory: ‘data/dogscats/sample/train’.

Here is my Cygwin window:

Any help would be hugely appreciated.



Howto: installation on Windows

Can’t verify this atm myself, but I think when you run a notebook the paths are relative to where the notebook resides and not relative to from where you run the jupyter notebook server.

You seem to have data under ~/nbs, but your ipython notebook is under courses/deeplearning1/nbs? In such a case, the relative path to the data folder would be: ‘…/…/…/data’ or something like that. You might want to move your data folder to the courses/deeplearning1/nbs folder and this should solve this for you.

To check what your notebok considers its working directory, you can do this:
import os

(Simon Laycock) #3

Thanks - moving the folder worked!

(Eyal) #4

I can’t find the dogscats.zip file in the course GitHub. Where should I get it ?

(Eyal) #5

I registered to kaggle and downloaded the two files. But, I could’t find sample zip file. Should we copy some images to the “sample” folder as we choose ?

(Jeremy Howard) #6

Look in http://files.fast.ai/data/

(Henry Marshall) #7

It would be really helpful if someone could add the new link was posted on the wiki’s lesson 1 and datasets pages. I’d do it myself, but it seems like you can’t edit the wiki as a MOOC student.

(Susan Li) #8

I am getting “Segmentation fault (core dumped)” error when downloading the dogscats.zip as follows, I have not found anyone asked this question. Can you please help? I am using aws p2.xlarge. Thank you !

(Alex Brie) #9

Hi, did you get any answer to it? I’m having the exact same issue as well :frowning:

(Alex Brie) #10

I replaced it with curl http://files.fast.ai/data/dogscats.zip -o dogscats.zip ;

I was also getting issues with unzip not being found, so I eventually fixed this latest one by doing:

sudo apt-get update
followed by
sudo apt install unzip

To get wget working, you can also call
sudo apt install wget

(Jeremy Howard) #11

Thanks for the suggestion - I’ve done that now.

(Jeremy Howard) #12

I think the problem occurs when wget is in too small a tmux pane to display its progress bar properly. Increasing the size of the tmux pane (e.g. by pressing ‘Ctrl-B’, then ‘z’) seems to fix it. Press the same keys again to return the pane to its original size.

(Susan Li) #13

Seems you have worked it out, I have not. Thanks for sharing,


Hello all, I think I have a similar problem, but I can’t seem to get a data directory in nbs:


I think this is why I’m getting the “OSError: [Errno 20] Not a directory: ‘data/dogscats/sample/train’” error when trying to execute vgg16

Thanks for the help!

(Allan Jackson) #15

Yes, Thanks, your suggestion to move the dogscats folder into the data file (in the same directory level as the notebooks files) worked great for me!