You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
On Windows (didn't test Linux) I was getting the error: Training data file "datasets\mnist\train-images-idx3-ubyte" not found; did you remember to download and extract it?: Os { code: 5, kind: permissiondenied, message: "access is denied." }
from running Mnist::new(r"datasets\mnist");
To get Mnist::new(r"datasets\mnist"); to work I had to move each of the idx3-ubyte files (from their folders with only the idx3-ubyte file inside after being extracted) into datasets\mnist\ and change the filenames const values in lib.rs to the following:
the only difference being a '.' instead of a '-' before 'idx3-ubyte' in each.
I'm not sure if I was doing anything wrong or if there is a better way for solving this. I used the built in file extraction tool in Windows 11 to decompress the .gz file, which might not be behaving how you expect (it created a folder with the name "train-images-idx3-ubyte" that has a file named "train-images.idx3-ubyte" in it from decompressing "train-images-idx3-ubyte.gz").
The text was updated successfully, but these errors were encountered:
Hmm thanks for reporting this, I'll check it out. I never tested this on windows 11, so could definitely be a quirk with some change in default behavior for windows 11 vs 10. Or it could be a change in the mnist archive upstream.
Just tested on NixOS/KDE Plasma, and the files from http://yann.lecun.org/exdb/mnist/index.html are extracted as expected. So this does seem to be a Windows quirk. I'll make a fix this weekend
On Windows (didn't test Linux) I was getting the error:
Training data file "datasets\mnist\train-images-idx3-ubyte" not found; did you remember to download and extract it?: Os { code: 5, kind: permissiondenied, message: "access is denied." }
from running Mnist::new(r"datasets\mnist");
To get Mnist::new(r"datasets\mnist"); to work I had to move each of the idx3-ubyte files (from their folders with only the idx3-ubyte file inside after being extracted) into datasets\mnist\ and change the filenames const values in lib.rs to the following:
the only difference being a '.' instead of a '-' before 'idx3-ubyte' in each.
I'm not sure if I was doing anything wrong or if there is a better way for solving this. I used the built in file extraction tool in Windows 11 to decompress the .gz file, which might not be behaving how you expect (it created a folder with the name "train-images-idx3-ubyte" that has a file named "train-images.idx3-ubyte" in it from decompressing "train-images-idx3-ubyte.gz").
The text was updated successfully, but these errors were encountered: