| Time |
S |
Nick |
Message |
| 02:46 |
|
prologic |
dotplus: lemme know when you try out golinks again :) hopefully I fixed the issues you ran in to :) |
| 09:41 |
|
|
semiosis joined #sourcefu |
| 09:41 |
|
|
semiosis joined #sourcefu |
| 13:27 |
|
pdurbin |
A screencast series of zombie-themed games written with Clojure and ClojureScript: http://www.parens-of-the-dead.com |
| 13:27 |
|
pdurbin |
Via https://defn.audio/2016/06/13/defn-episode-03-tour-of-the-repls/ |
| 13:48 |
|
|
dotplus joined #sourcefu |
| 13:48 |
|
|
dotplus joined #sourcefu |
| 17:48 |
|
pdurbin |
If anyone is interested in TensorFlow, here's the workshop I'm taking today and tomorrow: https://github.com/andreaazzini/DL_computefest_2017 |
| 20:10 |
|
pdurbin |
they seem to assume I know what "gradient descent" means |
| 20:10 |
|
* pdurbin |
looks at https://en.wikipedia.org/wiki/Gradient_descent |
| 20:11 |
|
pdurbin |
These are the slides for day 1: https://github.com/andreaazzini/DL_computefest_2017/blob/d46c949e58e64792c1c41e76793f028c5c4c3a4b/Slides/i.pdf |
| 20:37 |
|
pdurbin |
operating on this dataset: MNIST handwritten digit database, Yann LeCun, Corinna Cortes and Chris Burges - http://yann.lecun.com/exdb/mnist/ |
| 20:37 |
|
pdurbin |
https://github.com/andreaazzini/DL_computefest_2017/blob/master/MNIST.ipynb |
| 20:39 |
|
pdurbin |
"The mnist variable contains a TensorFlow high level Datasets object, handling all the operations we'll perform on the data (labeling images, counting iterations during training, etc)." |
| 20:45 |
|
pdurbin |
"As a first simple model, we are going to implement a neural network performing softmax regression, which is a straightforward way to do classification." |
| 20:50 |
|
pdurbin |
"The operation graph we define is just a flow of tensors, hence the name TensorFlow." |
| 21:28 |
|
pdurbin |
his solution to exercise 1: https://github.com/andreaazzini/DL_computefest_2017/blob/master/Ex1.ipynb |
| 21:33 |
|
pdurbin |
now on to https://github.com/andreaazzini/DL_computefest_2017/blob/master/TensorBoard.ipynb |
| 21:35 |
|
pdurbin |
"TensorBoard, which is a suite of visualization tools easily pluggable into your TensorFlow model" |
| 21:45 |
|
pdurbin |
tensorboard --logdir=summary |
| 22:02 |
|
pdurbin |
... but don't forget to run all cells in TensorBoard.ipynb to create the "summary" directory of tfevents |
| 22:18 |
|
pdurbin |
he keeps saying "cost function" but it's not really in the slides |
| 22:19 |
|
pdurbin |
I guess in the slides it's called a "loss function" (I think). |
