daviddao

//Job Result Schema
{
  "$id": "http://example.com/example.json",
  "type": "object",
  "definitions": {},
  "$schema": "http://json-schema.org/draft-07/schema#",
  "properties": {
    "job_id": {
      "$id": "/properties/job_id",
      "type": "string",

daviddao

def train(epoch):
    model.train()
    model.cuda()
    optimizer = optim.SGD(model.parameters(), lr=args.lr, momentum=args.momentum)
    if dist.get_rank() == 0:
        input_from_part2 = torch.FloatTensor(data.size()[0], 320)
        for batch_idx, (data, target) in enumerate(train_loader):
            optimizer.zero_grad()
            data = Variable(data.cuda())
            output = model(data)

daviddao

How to install Intel SGX, big picture

0. Check if your machine has Intel SGX (https://github.com/ayeks/SGX-hardware)
1. Enable it in BIOS (On Ubuntu, restart and press F2)
2. Install the SGX driver (https://github.com/01org/linux-sgx-driver), take care to switch to branch SGX2
3. Install Intel SGX SDK (https://github.com/01org/linux-sgx/). This allows you to compile programs in Simulation mode with flag make SGX_MODE=SIM
4. Install Intel SGX PSW. For that you need to download the capability licensing client from the Intel website! (https://software.intel.com/en-us/sgx-sdk/download)
5. If you make an SGX program, remember to always source environment!

That's it

daviddao

Transforming Code into Beautiful, Idiomatic Python

Notes from Raymond Hettinger's talk at pycon US 2013 video, slides.

The code examples and direct quotes are all from Raymond's talk. I've reproduced them here for my own edification and the hopes that others will find them as handy as I have!

Looping over a range of numbers

for i in [0, 1, 2, 3, 4, 5]:

daviddao

# Confusion Matrix (improved version with total numbers)
def plot_confusion_matrix(conf_arr, labels, title='Confusion matrix', cmap=plt.cm.Blues):
    #plt.imshow(cm, interpolation='nearest', cmap=cmap)
    norm_conf = []
    for i in conf_arr:
        a = 0
        tmp_arr = []
        a = sum(i, 0)
        for j in i:
            tmp_arr.append(float(j)/float(a))

daviddao

'''
Takes an image in numpy format as input and returns a padded image
'''
def pad_image(image, output_shape):
    x_pad = (output_shape[0] - image.shape[0]) / 2
    y_pad = (output_shape[1] - image.shape[1]) / 2

    resized = np.pad(image, ((x_pad,x_pad),(y_pad,y_pad)), mode='constant', constant_values=0)
    return resized

daviddao

import tensorflow as tf

def transformer(U, theta, downsample_factor=1, name='SpatialTransformer', **kwargs):
    """Spatial Transformer Layer
    
    Implements a spatial transformer layer as described in [1]_.
    Based on [2]_ and edited by David Dao for Tensorflow.
    
    Parameters
    ----------