skrish13

• Removing conda safely
  • rm -rf /path/to/anaconda/
  • vim ~/.bashrc
  • #comment out or delete - the line where conda was added to path
  • source ~/.bashrc
  • #now enjoy! :)

skrish13

 # Clean workspace
> rm(list=ls())
> 
> # Load MXNet
> require(mxnet)
> 
> # Loading data and set up
> #-------------------------------------------------------------------------------
> 
> # Load train and test datasets

skrish13

List of to-do's after installing GNOME:

• Extensions
  • Dash to Dock
  • Pixel Saver
  • Places indicator
  • Workspace indicator
  • Suspend icon in Menu
• Extra Apps
  • VLC

• Thunderbird

skrish13

• Set a particular brightness on startup

  • http://askubuntu.com/questions/151651/brightness-is-reset-to-maximum-on-every-restart
  • write in /etc/rc.local , echo desired_value > /sys/class/backlight/intel_backlight/brightness

• Bash shortcuts

  • Ctrl+R - reverse search lookup, history and greping.

skrish13

[nvcc]
optimizer_including=cudnn

[cuda]
root = /usr/local/cuda/bin

[global]
floatX = float32
device = gpu0
exception_verbosity = high

skrish13

• Installing CUDA

  • Most worked method - .deb file (of your distro) from the site
  • Installing it on a tty
  • Has problems with Mint as such

• Setting CUDA Paths

  • write in .bashrc:
    • #CUDA Paths
    • export PATH=/usr/local/cuda-8.0/bin:$PATH

• export LD_LIBRARY_PATH=/usr/local/cuda-8.0/lib64:/usr/local/cuda-8.0/lib

skrish13

from theano import function, config, shared, sandbox
import theano.tensor as T
import numpy
import time

vlen = 10 * 30 * 768  # 10 x #cores x # threads per core
iters = 1000

rng = numpy.random.RandomState(22)
x = shared(numpy.asarray(rng.rand(vlen), config.floatX))

skrish13

LO Papers:

• Old papers (not maintained 4r long time) - https://github.com/robertsdionne/neural-network-papers
• Few selected papers on each topic of DL - http://deeplearning.net/reading-list/
• Vision papers wrt each task - https://github.com/kjw0612/awesome-deep-vision

LO Conference Vids:

• https://github.com/dustinvtran/ml-videos

LO Tuts,projs etc:

• Tuts,projects,communities - https://github.com/ChristosChristofidis/awesome-deep-learning

skrish13

List of datasets by TUT:

Datasets released for the DCASE16 Workshop

• All datasets are only for validation and testing. Training is only available for task2-dcase2016

• TUT SE16 http://www.cs.tut.fi/sgn/arg/dcase2016/task-acoustic-scene-classification TUT Sounds Event 2016 DCASE16 - Task1

skrish13

import shutil, torch
def save_checkpoint(state, is_best, filename='checkpoint.pth.tar'):
    torch.save(state, filename)
    if is_best:
        shutil.copyfile(filename, 'model_best.pth.tar')
        
save_checkpoint({
    'epoch': epoch + 1,
    'arch': args.arch,
    'state_dict': model.state_dict(),