footer: @parisba / @the_mcjones / @themartianlife / #TFWorld theme: Zurich,6 slidenumbers: true
#[fit] 👋 Hello!
^ Hello!
[.build-lists: true]
git clone https://github.com/google/swift-jupyter.gitdocker build -f docker/Dockerfile -t swift-jupyter .docker run -p 8888:8888 --cap-add SYS_PTRACE -v /my/host/notebooks:/notebooks swift-jupyter
... so much Docker.
^... or use Google Colaboratory!
- Colab is a free, experimental data-science platform from Google
- It's basically a customised version of Jupyter Notebooks
- It's primarily Python
- but the Swift for TensorFlow team appears to have bolted Swift into it as well!
- it's a bit experimental, though
^ Originally, we weren't planning on using Colab for this session, but after spending a few weeks building the content we realised that, unless you have a truly powerful machine running Jupiter and the Swift Kernel locally, it really takes longer than is useful in a tutorial-setting to run any of the larger machine learning models.
- Activity 1: Making sure everything is working
- Activity 2: Getting familiar with Swift
- Activity 3: Slightly more advanced Swift
- Activity 4: Meet TensorFlow — Training a Model
- Activity 5: Swift and Python
- Activity 6: Building a GAN
- Activity 7: A little more Python
^ Throughout this plan, we're gonna be doing some activities, because, well, this is a tutorial! Our activities will be.... [advances one by one]. We may, or may not, get all the way through this.
- 11:50 AM tomorrow
- Great American Ballroom J/K
- performance!
- safety
- ease of reading
- fun
- looks kind of like Python most of the time
- performance
- performance
def collatz(n):
count = 0
while n > 1:
if n % 2 == 0:
n /= 2
else:
n = 3 * n + 1
count += 1
return count
print(f"it took {collatz(9)} steps to resolve")func collatz(number: Int) -> Int {
var count = 0
var n = number
while n > 1 {
if n % 2 == 0 {
n /= 2
} else {
n = 3 * n + 1
}
count += 1
}
return count
}
print("it took \(collatz(number: 9)) steps to resolve")| Python | Swift |
|---|---|
| 35.27 seconds | 0.88 seconds |
^ py had a stddev of 2.6s swift had a stddev of 0.13s so a ~40x performance
| Python | Swift | C |
|---|---|---|
| 35.27 seconds | 0.88 seconds | 0.0044 seconds |
^ I was curious so I rewrote it in C knowing it would be faster and was then confused why C was THAT much faster
^ so instead of reinterpreting Swift each time I compiled it once and ran that so it is an apples to apples for Swift and C but not for poor Python Swift is winning here but I only did 100 tests no where NEAR enough and we are talking about 4 nanoseconds here worth saying this isn't a slight at Python, it was designed for writeability first
| Python | Swift | C | Swift compiled |
|---|---|---|---|
| 35.27 seconds | 0.88 seconds | 0.0044 seconds | 0.0040 seconds |
^ We're going to be teaching Swift, as a programming language, for the first bit of this tutorial. We're kind of just ignoring machine learning for this. It's important to learn Swift as a programming language before we move to Swift for TensorFlow.
-
Go to this URL: