Author here!
It's a simple example but I was delighted to see how clean the code looked when implemented with outlines so I wanted to share it.
Feel free to ask questions!
114 karma · joined September 18, 2012
Personal website: https://morgangiraud.com
[ my public key: https://keybase.io/morgangiraud; my proof: https://keybase.io/morgangiraud/sigs/n_HlzVZBa-04X2SUaa4pPyufe3iGsKhSIKnMAm8clwI ]
It's a simple example but I was delighted to see how clean the code looked when implemented with outlines so I wanted to share it.
Feel free to ask questions!
It's a simple example but I was delighted to see how clean the code looked when implemented with outlines so I wanted to share it.
Feel free to ask questions!
Location: France, CEST timezone (UTC +2)
Remote: Yes
Willing to relocate: No
Technologies: Python, CUDA, pytorch/JAX/TF, Aws/Docker/Linux, typescript/React
Résumé/CV: https://morgangiraud.com/public/cv-morgan-giraud.pdf
Email: contact[AT]morgangiraud[DOT]com
More ML/Web projects at https://morgangiraud.com
Location: Annecy, France Remote: yes (hybrid) Willing to relocate: Maybe Technologies: Python, Pytorch/jax, Typescript, docker Résumé/CV: https://morgangiraud.com/public/cv-morgan-giraud.pdf Email: contact@morgangiraud.com
> How to speak to insufferable people?
When faced with insufferable people, it's important to focus on yourself and your own conduct. As Seneca said, "It is your conduct that you must hold in check; you must rouse what is languid in you, bind fast what has become relaxed, conquer what is obstinate, persecute your appetites, and the appetites of mankind, as much as you can".
You should also close your ears against evil talk and avoid being drawn into conversations that could lead you astray. As Seneca said, "Virtue, Philosophy, Justice this is a jargon of empty words. The only way to be happy is to do yourself well. To eat, drink, and spend your money is the only real life, the only way to remind yourself that you are mortal"
- I list all the different tasks one will have to do when doing ML
- I show a common folder structure that I believe handles all possible use cases nicely
- I show a basic Model class, easily extendable that structure a lot of possible kinds of models
- I describe how to build a good "shell API" for easy iterations.
Bonus: Some TF code linked to the subject.
- I explore the different ways to manually mutate Variables in TF (content and shape)
- I explore how to construct control flow like an "if statement"
- I end up and showing a weird TF behaviour when you mix those
- Bonus: a first try at animated GIF :D
I clearly don't expect people to do that much. I can only do that because i'm coming from web development, and very nice tools started to appear recently.
People in research needs a design framework like a set of templates for keynotes/PPT/JS/CSS (think about how much traction got bootstrap). Distill is doing an awesome jobs at showing the example of what you could do.
Maybe Distill could open-source the templates they use to build those blog post?
I've just spent some time learning TF on my own using the official doc and made a lot of projects on my Githubs.
If you're interested, here is the complete list of my articles so far:
How to handle shapes in TensorFlow: https://blog.metaflow.fr/shapes-and-dynamic-dimensions-in-te...
TensorFlow saving/restoring and mixing multiple models: https://medium.com/@morgangiraud/tensorflow-saving-restoring...
How to freeze a model and serve it with a python API: https://blog.metaflow.fr/tensorflow-how-to-freeze-a-model-an...
TensorFlow howto: a universal approximator inside a neural net: https://blog.metaflow.fr/tensorflow-howto-a-universal-approx...
How to optimise your input pipeline with queues and multi-threading: https://blog.metaflow.fr/tensorflow-how-to-optimise-your-inp...
https://blog.metaflow.fr/tensorflow-a-primer-4b3fa0978be3#.2...
I hope it can be useful for future learners! (you can find the current list of articles at the end of the first one)
TLDR:
- I implement a first basic neural net using the "feed_dict" system
- I show step by step how to use queues in TF
- I update the first example with a queue implementation and show that i train 33% faster the neural net
- bonus: Some more code on queues on my Github in the references section
Here is my new article on universal approximation theorem
TLDR: - I implement a first universal approximator with TesnorFlow and i train it on a sinus function (I show that it actually works) - I use it inside a bigger neural networks to classify the MNIST dataset - I display the learnt activation functions - I show that whatever the learnt activation function is, i get consistently the accuracy 0.98 on the test set - bonus: all the code is open-source
Feel free to ask questions or give feedback or both!
I finally find a solution with docker for mac to have an easy way to plug different virtual host to my Nginx on OSX dynamically.
Feedback welcome!
"Less Clutter block" Our only focus is search. That means that our results are free of promotoions <--
"Discuss block" Our forums are a place -->for to<-- chat