ilblackdragon

import sys
import traceback


def _format_value(key, value, first_n=20, last_n=20):
    s = repr(value)
    s = s.replace('\n', ' ').strip()
    if len(s) > first_n + last_n + 3:
        s = s[:first_n] + "..." + s[-last_n:]
    return "%s: %s" % (key, s)

ksopyla

# This is shorthened version of blog post 
#  http://ksopyla.com/2017/02/tensorflow-gpu-virtualenv-python3/

# update packages
sudo apt-get update
sudo apt-get upgrade

#Add the ppa repo for NVIDIA graphics driver
sudo add-apt-repository ppa:graphics-drivers/ppa
sudo apt-get update

iMilnb

#!/usr/bin/env python

# Simple [boto3](https://github.com/boto/boto3) based EC2 manipulation tool
#
# To start an instance, create a yaml file with the following format:
#
# frankfurt:
# - subnet-azb:
#   - type: t2.micro
#     image: image-tagname

eamartin

harpone

import tensorflow as tf
from tensorflow.python.framework import ops
import numpy as np

# Define custom py_func which takes also a grad op as argument:
def py_func(func, inp, Tout, stateful=True, name=None, grad=None):
    
    # Need to generate a unique name to avoid duplicates:
    rnd_name = 'PyFuncGrad' + str(np.random.randint(0, 1E+8))
    

apaszke

import torch                                                                                          
                                                                                                      
def jacobian(y, x, create_graph=False):                                                               
    jac = []                                                                                          
    flat_y = y.reshape(-1)                                                                            
    grad_y = torch.zeros_like(flat_y)                                                                 
    for i in range(len(flat_y)):                                                                      
        grad_y[i] = 1.                                                                                
        grad_x, = torch.autograd.grad(flat_y, x, grad_y, retain_graph=True, create_graph=create_graph)
        jac.append(grad_x.reshape(x.shape))