b2net / AI_DL_AWS

materials around AI, Deep learning, AWS ...
0 stars 0 forks source link

"descriptive statistics" der Zalando Bilddaten #1

Open lborke opened 6 years ago

lborke commented 6 years ago

Auch zum Datensatz: was ist genau drin, etwas "descriptive statistics" und Hintergrundinfo zu den Daten, wie viele, welche Labels, etc.?

lborke commented 6 years ago

Du kannst dazu gerne eine neue Python-Datei mit notwendigen Funktionen etc. anlegen. Ich bin sicher, dass Numpy Pandas Sklearn etc. sinnvoll sein werden.

anastasia-stepanchenko commented 6 years ago

https://github.com/lborke/AI_DL_AWS/blob/master/MXNet/Fashion_MNIST_Data_Description.ipynb

Вот то, что пока в голову пришло.. Буду добавлять по мере поступления мыслей.

lborke commented 6 years ago

Hier ist meine Version, die deine korrigiert, verbessert, und "richtig" formatiert: https://github.com/lborke/AI_DL_AWS/blob/master/MXNet/Fashion_MNIST_Data_Description_LB.py

anastasia-stepanchenko commented 6 years ago

https://github.com/lborke/AI_DL_AWS/blob/master/MXNet/Fashion_MNIST_Data_Description.ipynb внесла твои изменения

Wie kann man das noch kompakter und eleganter machen. De facto wird hier ein Loop gemacht. [fashion_labels[i] for i in train_lbl[0:pict_number].tolist()]
заменила на operator.itemgetter(*train_lbl[0:pict_number])(fashion_labels) везде пишут, что она быстрее