image-classification

Enhancement

A discussion in #614 revealed a good place for improvement - we should ensure that input image is continuous upon start of the augmentation pipeline. This could be implemented by adding image = np.ascontiguousarray(image) to image and mask targets.

A proposed place to add this call - somewhere at the beginning of A.Compose.__call__.

Steps:

Register on the website
An account verification link will be sent by email
Go to email inbox
Right-click on the box and copy the link
Check the link by pasting it in new tab
Hit enter and check if the account is being opened

HTTP LINK:

Impact - The user's account can be directly logged if the token is grabbed by the attacker by the tool named Wireshark.

**I want to

There are many links in Kinetics that have expired. As as result, everyone might not be using the same Kinetics dataset. As a reference, the statistics of the Kinetics dataset used in PySlowFast can be found here, https://github.com/facebookresearch/video-nonlocal-net/blob/master/DATASET.md. However, I cannot seem to find similar information for gluoncv. Will you guys be sharing the statistics and

When running TabularPredictor.fit(), I encounter a BrokenPipeError for some reason.
What is causing this?
Could it be due to OOM error?

Fitting model: XGBoost ...
-34.1179 = Validation root_mean_squared_error score
10.58s = Training runtime
0.03s = Validation runtime
Fitting model: NeuralNetMXNet ...
-34.2849 = Validation root_mean_squared_error score
43.63s =

with video_demo.py about 20% speed compared to your 1.0 repo. but thanks much for sharing!