Profile picture
François Chollet
@fchollet
, 4 tweets, 1 min read Read on Twitter
We will rebuild -- but it will take decades.

I'm from Nantes, a city with a beautiful medieval cathedral. Its roof was entirely destroyed in a fire in 1972, in very similar circumstances. It took 40 years to rebuild.
The good news is that nothing truly irreplaceable was lost. Cathedrals get perpetually rebuilt as part of their regular lifecycle - like the ship of Theseus. The spire that fell was a 1860 addition by Viollet-le-Duc.

It's normally meant to be a more incremental process though...
Fun fact: when it comes to the Nantes cathedral, the 1972 fire that destroyed the roof was started by a roofer with a blowtorch, working on the finishing touches of restoration work that was addressing damage from... the 1944 allied bombings. Work is never over.
It took 13 years to rebuild the roof itself, but the complete restoration work only ended in 2008.
Missing some Tweet in this thread?
You can try to force a refresh.

Like this thread? Get email updates or save it to PDF!

Subscribe to François Chollet
Profile picture

Get real-time email alerts when new unrolls are available from this author!

This content may be removed anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

More from @fchollet see all

Profile picture
François Chollet
@fchollet
This is how you implement a network in Chainer. Chainer, the original eager-first deep learning framework, has had this API since launch, in mid-2015.

When PyTorch got started, it followed the Chainer template (in fact, the prototype of PyTorch was literally a fork of Chainer).
Nearly every day, I am getting ignorant messages saying, "PyTorch is an original innovation that TensorFlow/Keras copied". This is incorrect. Subclassing is a fairly obvious way to do things in Python, and Chainer had this API first. Many others followed.
I had been looking at adding a Model subclassing API to Keras as soon as late 2015 (before the Functional API even existed, and over a year before being aware of PyTorch), inspired by Chainer. Our first discussions about adding an eager execution mode also predate PyTorch.
Read 6 tweets
Profile picture
François Chollet
@fchollet
When you're using tf.keras in TensorFlow 2.0, training your models with `fit` is one option, but you can also write a custom training loop. It takes about 10 lines.
If your model creates additional losses during its forward pass (i.e. if you called `self.add_loss` in a custom layer), collect them in one line like this.
The advantage of using `fit` is that it packs a lot of functionality in a convenient interface: a progress bar, easy reporting of metrics such as precision & recall, easy management of class weights for imbalanced classification, callbacks for model checkpointing and more...
Read 3 tweets
Profile picture
François Chollet
@fchollet
Just like everything else, thinking takes practice.
Thinking is recombining largely preexisting concepts following largely preexisting patterns. Infrequently do we generate brand new ones -- it takes a lot of effort. As such, deep thinking requires practice in order to accumulate a dense network of these concepts and patterns.
Of course, it also requires to read books. It would be challenging to have to recreate thousands of years worth of accumulated thought building blocks on your own.
Read 4 tweets

Related threads

Profile picture
Adrianna Tan
@skinnylatte
To follow up on yesterday’s food tweet storm I’d like to post about some of my fave SE Asia travel and food experiences!
Being from Singapore I was always into short trips in my region and I’ve been going to Malaysia for as long as I was conscious and alive to visit distant family, but mostlyto eat. Eventually I lived in KL for several years, too. I also got to see Malaysian Borneo.
I’m most familiar with KL and JB because that’s where I’ve spent most of my time. My fave KL foods: chilli pan mee, broad chewy udon-like homemade noodles tossed with homemade spicy chilli that’s an umami bomb. I like Kin Kin and Super Kitchen for this dish.
Read 33 tweets
Profile picture
Lars Pellinat
@Lars9596
1. #NotreDame: One of the most magical places on earth.

The Cathedral of Notre-Dame de Paris was built on a site which in Roman Lutetia is believed to have been occupied by a pagan temple.

#Paris #Fire #NotreDame
2. #NotreDame: The beautiful #Spire of the Cathedral, originally in place 13th-18th century, recreated in the 19th century, has now been completely destroyed by the flames (see middle of the picture).
#PARIS
3. #NotreDame:
- In 1163 the Cornerstone was laid for Notre-Dame de Paris and construction started.
- In 1250 Western towers and north #RoseWindow completed.
- In the 14th century most of the Cathedral was completed.

Picture: Rose Window
Read 30 tweets
Profile picture
mugrimm
@unabanned
Going to take this opportunity to remind people that Silicon Valley paved the way for this narrative by presenting an entire class of people with very similar backgrounds as totally 'self made' despite the fact it's pretty clear they got where they got because of their parents.
Almost every silicon valley creator comes from the exact same background. Usually upper middle class parents with at least one parent having an advanced degree, usually in math/coding/electrical engineering, or similar field. They got big because they literally had no competition
Larry Page, one of the founders of Google and CEO of alphabet: Father was a professor of computer science in the 70's, mother taught coding. Both upper middle class.
Read 20 tweets
Profile picture
Nuddering
@NudderingNudnik
Thread on @PeoplesMomentum and antisemitism:

This week Momentum released a video on antisemitism that featured a young Jewish woman who made a number of very problematic comments.
Good analyses have been provided by @daverich1 @DavidHirsh @mishtal and @israel_advocacy
The video represents just another example of the way that the Left highlights and celebrates Jewish voices who represent only a very small minority of British Jews and yet pillories, abuses and condemns those Jewish voices that speak for the vast majority

The mainstreaming and empowering of the fringe, crank and wholly unrepresentative voice of those behind Jewish Voice for Labour and Free Speech on Israel is the most obvious and dangerous element of this current phenomena.

And we can see something very similar in another context
Read 11 tweets
Profile picture
Peter Kaufman
@PeterKaufmanSOC
Beautiful
(A microburst Twitter essay--to be finished before I'm done with my cough drop)

I just had an epiphany: #beautiful has become my new favorite word. That may seem strange given my circumstances. In June 2017 I was diagnosed with incurable stage IV #lungcancer.
I naively thought that maybe my lifetime of being a health and fitness junkie would insulate me from such a medical fate. Wrong. And So here I am, fifty-one years old and not sure how much time I have left on earth. I know I could, or maybe should, feel angry, cheated, jaded.
But I don't. I have this epiphany that my new favorite word is beautiful and I'm not sure why. There is nothing beautiful about what I've gone through since my diagnosis--lots of physical and emotional distress to be sure. But yet, I keep find
Read 15 tweets
Profile picture
Darshg9
@Darshg9
#Nantes Bon, essayons d'apporter un peu de bon sens dans les réseaux sociaux (oui, je crois au Père Noël).

Tout d'abord, sachez que j'habite au #Breil, juste à côté du lieu du drame, que je suis un père de famille sans histoires, avec un travail et qui essaye de faire en sorte
que sa femme et son petit ne manquent de rien. Depuis la mort d' #aboubakar, j'entends toutes les nuits des émeutes avec lesquelles je n'ai rien à voir et je vois le matin les dégâts des incendies, tout ça tout autour de mon immeuble.

Donc déjà, tous ceux qui commentent la
situation du quartier sans y vivre je leur répond: sois gentil et tais toi, tu sais pas de quoi tu parles et même à toi je ne te souhaite pas de vivre ce que je vis depuis le début de la semaine.

Ensuite, ceux qui se permettent de juger les faits, peu importe leur camp, qui
Read 32 tweets

Trending hashtags

#onhere #1A #FinancialIndependence #UnitedStates #FelizDomingo #cancer #War #neocon #HandsOffVenezuela #organs #London #HousingforHealth #diplomatic #RecordStoreDay2019 #rebuild #take #HumanRights #TrumpShutdown #SAA #PEoH #Destroyed #UNCharter #KhashoggiMurder #Freedom #BonSens
Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!