There are lots of little kiddy programmes I don't get to see now my youngest is at school every day, Mister Maker was one of my favourites though - utterly bonkers, lots of fun and we've made lots of the things featured on the show. I haven't watched Mister Maker for a g e s. Click HERE to make it over at Debenhams.We can load up some random chairs and see how different players drew chairs from around the world.Debenhams have a rather lovely deal on these rather fabulous Mister Maker Doodle Drawers - they're £10.50 now instead of £15 The Facets team has even taken the liberty of hosting it online and giving us some presets to play around with! You can access the page here. If you want to explore the dataset some more, you can visualize the quickdraw dataset using Facets. Maybe only do it for a subset of the data the first time around, on account of training time :) Data exploration and visualization of Quick Draw In a wonderous turn of events, there’s a guide specifically for using RNNs on the Quick Draw dataset, so check out the tutorial if you are interested in trying that out. If you want to be fancy and use the full dataset (fair warning, it’s pretty large!), you’ll likely want to use a Recurrent Neural Network (RNN) to get the job done, since it will learn from the sequence of strokes drawn.

So if you’re looking for something fancier than 10 handwritten digits, you can try processing over 300 different classes of doodles. Why is it 28x28? Well, it’s a perfect replacement for any existing code you might have for processing MNIST data. npy format, which can be loaded using np.load().

The fourth format takes the simplified data and renders it into a 28x28 grayscale bitmap in numpy.

The doodle drawer