|
Rick Wierenga 👨🏼💻
@
rickwierenga
The Netherlands 🇳🇱
|
|
Learning and I want my computer to do the same | WWDC ‘19 | 16 y/o | author @fritzlabs | my own blog at rickwierenga.com
|
|
|
491
Tweetovi
|
245
Pratim
|
941
Osobe koje vas prate
|
| Tweetovi |
| Rick Wierenga 👨🏼💻 proslijedio/la je tweet | ||
|
MIT CSAIL
@MIT_CSAIL
|
5 h |
|
MIT's new CS class teaches you things that all the other classes don't teach you, like...
🖥️Shell tools and scripting
🖥️Vim
🖥️Data wrangling
🖥️Command-line environment
🖥️Version control
Watch all 11 lectures for free here: bit.ly/MissSemester
#MondayMotivation pic.twitter.com/vHUkzqbXT7
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
8 h |
|
Hi Elon, do you offer summer internships?
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
2. velj |
|
A scholarship would also do 🙃
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
2. velj |
|
I agree. Open to offers 😄
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
2. velj |
|
Would that mean being able to train on macOS? That could get a lot more people (iOS devs) into the field and make a huge difference.
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
2. velj |
|
Ah great!
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
import pandas as 🐸
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
Getting rejected by @gvanrossum himself: github.com/python/cpython… lol
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
import tensorflow as 🐼*
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
Should I turn "An introduction to machine learning through polynomial regression" into a series called "ML from Scratch" series where I discuss the mathematics behind ML foundations?
rickwierenga.com/blog/machine%2…
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
Honestly that's awesome!
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
Just thinking here, would it be possible to train on device? @mhollemans wrote a great series on how to do it on iOS.
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
I'm not sure about your specific use case, but do your models have to be retrained from scratch every time? You could consider incremental training which is much cheaper.
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
Oh that's a lot. But isn't training it locally on a Mac even slower?
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
TFLite models work great on iOS *and* Android, btw. But they can only be trained in Python TensorFlow (for now).
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
I'm primarily a Mac user myself but use Google Cloud or Colab for most things ML. (Swift is supported by Colab too!) It works great for training almost all models, certainly the ones used on mobile.
I hope Swift gets TFLite support soon :) (cc @bradlarson?)
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
Why wouldn't you use the best software (in this case) available? Not using CUDA because other people use it seems odd.
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
Check out Swift for TensorFlow!
tensorflow.org/swift
|
||
|
|
||
|
Rick Wierenga 👨🏼💻
@rickwierenga
|
1. velj |
|
how to mess with your coworkers pic.twitter.com/l7qnB6Kx5p
|
||
|
|
||