Skip to content

Commit cadad3e

Browse files
dnikudniku
committed
Remove theano/lasagne
1 parent b626bdd commit cadad3e

24 files changed

+38
-7241
lines changed

‎Amazon GPU howto.md

+7-8
Original file line numberDiff line numberDiff line change
@@ -1,22 +1,21 @@
1-
# How to set up GPU on EC2 instance
1+
**Warning: this guide has not been updated since we migrated away from Theano/Lasagne, which happened a long time ago.**
22

3-
## Create EC2 instance
3+
# How to set up an Amazon EC2 GPU instance
4+
5+
## Create an instance
46

57
Use `p2.xlarge` instance type and `ami-e00a8180` AMI image. [Details](http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html)
68

7-
Open ports `22` (ssh) and `80` (http) on your freshly created instance,
8-
you create a [security group](http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/using-network-security.html)
9-
and attach it your instance to get ports open
9+
Open ports `22` (ssh) and `80` (http) on your freshly created instance, then create a [security group](http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/using-network-security.html) and attach it to your instance to get the ports open.
1010

1111
## Launch notebook
1212

13-
Instance you have created contains all you need: fresh versions of theano, lasagne, CUDA driver and cuDNN,
14-
just lunch ipython and get hands dirty:
13+
The instance you have created contains all you need: fresh versions of theano, lasagne, CUDA driver and cuDNN,
14+
just launch Jupyter and get your hands dirty:
1515

1616
```bash
1717
$ sudo su
1818
$ export THEANO_FLAGS='cuda.root=/usr/local/cuda,device=gpu,floatX=float32'
1919
$ export PATH=/usr/local/cuda-8.0/bin${PATH:+:${PATH}}
2020
$ jupyter notebook
2121
```
22-

‎week04_[recap]_deep_learning/README.md

+8-12
Original file line numberDiff line numberDiff line change
@@ -1,22 +1,20 @@
1-
__Note:__ This week's materials cover the basics of neural nets and deep learning and teach you how to use auto-diff frameworks. If you're already fluent in tensorflow OR pytorch OR theano - feel free to skip this week entirely..
1+
__Note:__ This week's materials cover the basics of neural nets and deep learning and teach you how to use auto-diff frameworks. If you're already fluent in Tensorflow or PyTorch, feel free to skip this week entirely.
22

33
## Materials
44
* [__Lecture slides__](https://yadi.sk/i/yAO2AJ3M3EKP8g)
55

66
- __In russian:__
77
* Basic lecture on deep learning - [video](https://yadi.sk/i/yyHZub6R3Ej5dV)
8-
* Deep learning frameworks - [video](https://yadi.sk/i/hDIkaR4H3EtnXM)
8+
* Deep learning frameworks - [video](https://yadi.sk/i/hDIkaR4H3EtnXM)
99
* [Pytorch tutorial](https://yadi.sk/i/O3mQ76u43So3h9) __recommended__
1010
* [Tensorflow tutorial](https://www.youtube.com/watch?v=FQ660T4uu7k) (english only for now. Links are welcome)
11-
* [Theano tutorial](https://yadi.sk/i/54STsEBVpubkn)
1211

1312
- __In english:__
1413
* Intro to neural nets and backprop (english) - [video](https://www.youtube.com/watch?v=uXt8qF2Zzfo)
1514
* Intro to convnets - [video](https://www.youtube.com/watch?v=FmpDIaiMIeA)
1615
* Deep learning frameworks - [video](https://www.youtube.com/watch?v=Vf_-OkqbwPo)
1716
* [Tensorflow tutorial](https://www.youtube.com/watch?v=FQ660T4uu7k)
18-
* [Theano tutorial](https://www.youtube.com/watch?v=OU8I1oJ9HhI)
19-
* [Pytorch tutorial](https://www.youtube.com/watch?v=VMcRWYEKmhw)
17+
* [PyTorch tutorial](https://www.youtube.com/watch?v=VMcRWYEKmhw)
2018

2119
## Bonus materials
2220
* Karpathy's course on deep learning (english) - http://cs231n.github.io/
@@ -26,21 +24,19 @@ __Note:__ This week's materials cover the basics of neural nets and deep learnin
2624
* Deep learning demystified - [video](https://www.youtube.com/watch?v=Q9Z20HCPnww)
2725
* Karpathy's lecture on deep learning for computer vision - https://www.youtube.com/watch?v=u6aEYuemt0M
2826
* Our humble DL course: [HSE'fall17](https://github.com/yandexdataschool/HSE_deeplearning), [Skoltech/YSDA'spring16](https://github.com/ddtm/dl-course/) courses on deep learning (english).
29-
* Srsly, just google `"deep learning %s"%s for s in what_you_want_to_know`.
27+
* Srsly, just google `"deep learning %s" % s for s in what_you_want_to_know`.
28+
3029

31-
3230
### Practice
33-
__[Colab url (pytorch)](https://colab.research.google.com/github/yandexdataschool/Practical_RL/blob/master/week04_%5Brecap%5D_deep_learning/seminar_pytorch.ipynb)__
34-
From now on, we'll have two tracks: theano and tensorflow. We'll also add pytorch seminars as soon as they're ready.
31+
__[Colab URL (PyTorch)](https://colab.research.google.com/github/yandexdataschool/Practical_RL/blob/master/week04_%5Brecap%5D_deep_learning/seminar_pytorch.ipynb)__
32+
From now on, we'll have two tracks: Tensorflow and PyTorch.
3533

36-
Please pick seminar_theano.ipynb, seminar_tensorflow.ipynb or seminar_pytorch.ipynb.
34+
Please pick `seminar_tensorflow.ipynb` or `seminar_pytorch.ipynb`.
3735

3836
__Note:__ in this and all following weeks you're only required to get through practice in _one_ of the frameworks. Looking into other alternatives is great for self-education but never mandatory.
3937

4038
#### What to choose?
4139
* The simplest choice is PyTorch: it's basically ye olde numpy with automatic gradients and a lot of pre-implemented DL stuff... except all the functions have different names.
4240
* If you want to be familiar with production-related stuff from day 1, choose TensorFlow. It's much more convenient to deploy (to non-python or to mobiles). The catch is that all those conveniences become inconveniences once you want to write something simple in jupyter.
43-
* Theano works like tensorflow but it offers a numpy-compatible interface and comes with built-in graph optimization. The payoff is that theano is not as popular as the first two. It is also not meant as a producton framework so deploying to mobiles may be a problem.
4441

4542
* It's not like choosing house at Hogwarts, you'll be able to switch between frameworks easily once you master the underlying principles.
46-

0 commit comments

Comments
 (0)