Saumitro Dasgupta

Oddly, I can't seem to reproduce it. I tested on Ubuntu 14.04 with both Python 2 and 3. Tensorflow installed using the `tensorflow-gpu` pip package: ```python Python 3.5.2 |Anaconda 4.2.0...

The reverse conversion is fairly similar: 1. Map TensorFlow ops (or groups of ops) to Caffe layers 2. Transform parameters to match Caffe's expected format Things are slightly trickier for...

The prototxt looks invalid - I'd expect the types to be quoted ([example](https://github.com/BVLC/caffe/blob/master/models/bvlc_reference_caffenet/deploy.prototxt)).

Currently, it's a bit tricky to cleanly export to PDF (besides doing something along the lines of Print -> Save as PDF...). You can try exporting the SVG using something...

In a bit of an abuse of convention, `npm start` currently launches the `watchify` process that generates `netscope.js`. However, once it has been generated, you can launch a local instance...

Netscope is written in [CoffeScript](http://coffeescript.org/). The [`watchify`](https://github.com/jnordberg/coffeeify) command basically watches for changes to the CoffeeScript sources and compiles them into a single javascript file ([`netscope.js`](https://github.com/ethereon/netscope/blob/gh-pages/assets/js/netscope.js)). The parser (generated from the...

## CUDA I believe the primary issue is that Python's `multiprocessing` defaults to `fork` (for Python < `3.14` on Linux), and the CUDA RT doesn't play well with that (eg:...