You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
TF tweetorial: if you have a list of paths (strings) to mp3 files on disk, and a list of labels (integers) for these files, here's how you'd make a labeled audio dataset (supporting buffering and prefetching, and robust against corrupted files) Note: tfio is tensorflow-iopic.twitter.com/xIFp91RIzG
Here we use only the first ~1 min of audio (1M timesteps -- mp3 is encoded at 16 kHz), and a single channel. The resulting data could be processed by a 1D convnet. You could also apply a FFT to it before sending it to your convnet, via `tf.signal.fft`.
You can alternate between the channels at train time. This is a really easy way to create augmentations :)
Interesting idea!
Lovin these Tweetorials, thank you.
@ariG23498 @snehangshu_ @Cr0wley_zz this is a cool pipeline. Our audio pipeline involved a lot of us.
This is really cool! 
To really understand how this works under the hood - checkout this excellent channel https://www.youtube.com/channel/UCZPFjMe1uRSirmSpznqvJfQ … by @musikalkemist
Thanks man! that's impressive!
@fchollet you are using the first minute of each audio file? When working with music, do you think this makes sense or is worth doing a sliding window or a previous segmentation?
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
