VARUNA JAYASIRI

@vpj

Generative Adversarial Networks (Gaussian Distribution)

February 3, 2018

I tried implementing a simple 1 dimensional Generative Adversarial Network (to fit a Gaussian distribution) on ObservableHQ.

I loved the observable HQ experience; deeplearn.js not so much.

The only annoying thing I came across with observable is that it started running all the cells at the startup. It sometimes causes the browser to stall when you load a notebook with heavy computation (may be observable is not meant for anything with heavy computation). The other problem with it was that if you have some animation running below the screen height, you will miss whatever happened before you scrolled down. May be they should let the user select which cells to run automatically, and only run those and their dependencies at startup.

I'm looking forward to using observable notebooks more.

Here's the notebook. Or on the iframe below.

I tried implementing a simple 1 dimensional Generative Adversarial Network (to fit a Gaussian distribution) on ObservableHQ. I loved the observable HQ experience; deeplearn.js not so much. The only annoying thing I came across with observable is that it started running all the cells at the startup. It sometimes causes the browser to stall when you load a notebook with heavy computation (may be observable is not meant for anything with heavy computation). The other problem with it was that if you have some animation running below the screen height, you will miss whatever happened before you scrolled down. May be they should let the user select which cells to run automatically, and only run those and their dependencies at startup. I'm looking forward to using observable notebooks more. <<https://beta.observablehq.com/@vpj/gan(Here's the notebook)>>. Or on the iframe below. <!> <<<html <script type="text/javascript"> function iframeLoaded() { var iframe = window.document.getElementById('numpy_lstm_iframe') if(iframe) { function setHeight() { iframe.height = iframe.contentWindow.document.body.scrollHeight + "px" } setTimeout(setHeight, 1000) setHeight() } } </script> <iframe id="numpy_lstm_iframe" onload="iframeLoaded()" src="https://beta.observablehq.com/@vpj/gan" style="width: 100%; border: none; outline: none; min-height: 640px;" />