Kalanit

Ph.D. Ecole Polytechnique
Deep Learning
Israel

[Kalanit @KalanityL]

@SoMabrouk

[Kalanit @KalanityL]

"Il y avait un avait un avant le 7 octobre quand même" ?!!
sur @CNEWS ? juste avant l'hommage ? que s'est-il passé @GWGoldnadel @PascalPraud @ELevyCauseur ??

[François Chollet @fchollet]

A neat TF2/Keras implementation of text-to-speech with Transformer. https://t.co/SBn9EyPf7T

[François Chollet @fchollet]

A neat implementation of YOLO v3 with TF 2.2 and Keras https://t.co/MZ7aTQAZ1K

[Gabriel Oliveira-Barra 🇦🇩 @nospotfer]

@fchollet Here's my favourite implementation in TF 2.x WITH SUPPORT FOR YOLO v4 with many different backbones, from efficientnet, mobilenetv3, xception, etc. It also provides automation for finding best hooks.
Plus, the code is constantly being maintained.
https://t.co/kQEld5PzIs

[François Chollet @fchollet]

A neat new workflow example on https://t.co/m6mT8SrKDD, by @dn_griffiths: an implementation of PointNet for 3D point cloud classification - https://t.co/iLm2hXTh0l

[François Chollet @fchollet]

New code example on https://t.co/m6mT8SrKDD: visualizing what convnets learn https://t.co/HsLzh7uNwc

[François Chollet @fchollet]

New example on https://t.co/m6mT8SrKDD: timeseries anomaly detection
https://t.co/pK2xRUsmta

[François Chollet @fchollet]

New example on https://t.co/m6mT8SrKDD: model interpretability with Integrated Gradients, by @A_K_Nain

https://t.co/zfxGvcKH2i

[Kalanit @KalanityL]

Useful #Python libraries for #DataScience tasks other than the commonly used ones like #pandas, #scikit-learn , #matplotlib, etc. https://t.co/GYWP5IMLZO

[Kalanit @KalanityL]

@VillaniCedric @mounir Docteur de l'Ecole Polytechnique, experte deep learning en Israel depuis 4 ans, j'envisage un retour en France. Il avait ete question d'aides pour inciter au retour en France les scientifiques. Est-ce concretise? En vous remerciant par avance, K.

[Kalanit @KalanityL]

#Notebook introduction to #PyTorch https://t.co/RvEOf79t3L Great work ! thanks to @joshfp

[Jeremy Howard @jeremyphoward]

The heart of a @PyTorch training loop with callbacks. By aligning the training code and callback code, you can see exactly what's going on in each.

Formatting code for understanding is too important to leave to automated tools or hard and fast rules.

[Kalanit @KalanityL]

#Pytorch #GAN #ModeleZOO https://t.co/zBoG487ctW via @ylecun

[Jeremy Howard @jeremyphoward]

This is so cool.

NB: it's not just for tensorflow! It's open source and any framework or language can use it 😀

[Jeremy Howard @jeremyphoward]

Resnets 18, 34, 50, 101, and 152, with all the tweaks from the "Bag of Tricks" paper (and more), in one screen of @pytorch code ij @ProjectJupyter .

Took two days of refactoring to get to this point, but now it's *so* easy to tweak and see exactly what's going on. :)

[ShubhamS @Tradest3]

Apparent color bias in word vectors. Observed while doing the first assignment of CS224N from @stanfordnlp. This important ethical inclusion in the assignments will make sure the future practitioners are wary of it. Kudos to @chrmanning and team. #NLP

[François Chollet @fchollet]

If you've been solving big problems with Keras + TensorFlow, I encourage you to submit a talk proposal to TensorFlow World, a new conference coming this October: https://t.co/abj0P0SwxT

[Kalanit @KalanityL]

@DeepMindAI &al teaching #arithmetic to #NeuralNetworks by representing numerical quantities as linear activations
https://t.co/ZpNftusXBC
#DeepLearning

[Jeremy Howard @jeremyphoward]

Up to 100% accuracy for cellular histo-pathology image classification… with Fast-AI and Kimia960… by Less Wright https://t.co/1KvqaJ6c1G