search

beckstev / MachineLearningSeminar

MachineLearningSeminar SS19 TU Dortmund
MIT License
0 stars 0 forks source link

Do we need a Dataloader? #4

Closed beckstev closed 5 years ago

beckstev commented 5 years ago

I am not sure if our RAM can handle ~15K images at once. So maybe we need a Dataloader which loads all images on the fly.

Check this out:

https://stanford.edu/~shervine/blog/keras-how-to-generate-data-on-the-fly

FeGeyer commented 5 years ago

In der ___datageneration-Methode müssen wir np.genfromtxt oder das äquivalent von Pandas nutzen, da man mit np.load keine csv Dateien laden kann.

beckstev commented 5 years ago

We can use pandas.read_csv('path_to_file') .

beckstev commented 5 years ago

Implemented