Deep metric learning helps to reduce the images we need for training on deep learning applications. Constellation loss code released: arxiv.org/abs/1905.10675 CCMIJU - JUMISC TECNALIA Horizon 2020 Tyndall Institute Karl Storz L4TECH bioef @M2Lasers #LENS Imperial College London

In this thread, I want to compile a list of Deep Learning resources 📚🎬 that some people might not be aware of. It’s amazing how organizations like alexandre, Stanford NLP Group and @openai and people like Rachel Thomas ,@pieterabbeel and Lex Fridman have made these freely available 🎉👇

Named Tensors! Built in to PyTorch 1.3. Now I won’t need a comment on every single line of math code documenting the Tensor dimensions. #ptdc19

We just released PyTorch3D, a new toolkit for researchers and engineers that’s fast and modular for 3D deep learning research: ai.facebook.com/blog/-introduc…

Si te molesta que los científicos hoy digan una cosa y mañana otra, ve y búscate un cura; ellos siempre han dicho, dicen y dirán lo mismo.

Our new paper (w/Karan Desai (KD)) argues that "language is all you need" for good visual features: we train CNN+Transformer *from scratch* on ~100k images+captions from COCO, transfer the CNN to 6 downstream vision tasks, and match/exceed ImageNet features despite using 10x fewer images!

Joan Serrà And so the era of implicit functions begins

PyTorch Lightning 1.0.0 is now available. This is the final stable API to train and deploy models at scale, without the boilerplate. Read more about this release below: medium.com/pytorch/pytorc…

Sarri Sarri - Kortatu feat. Boston Dynamics 🎶

Your front-row seat to my Mars landing is here. Watch how we did it. #CountdownToMars

fuzzcon.eu/2021/?utm_sour…

Halloween costumes to scare scientists (cartoon for New Scientist ) #halloween #AcademicTwitter

Euskara ulertzen duten Adimen Artifizialeko sistemak sortzeko beti izan dugu traba handi bat: kalitatezko corpus publiko baten falta. Hutsune hori betetzeko asmoz sortu dugu EusCrawl. Azalpen gehiago eta gogoeta batzuk jarraian.

Second order effects of the rise of large language models:

📢📢On Tuesday! Aitor Alvarez-Gila is defending his PhD "Self-supervised learning for image-to-image translation in the small data regime"🎓 📆19 July 11 am 📍 #CVC Conference room & Teams Directors: Dr Joost van de Weijer & Dr Estibaliz Garrote (@Tecnalia) #CVCPeople #phdlife

I am happy to share that today I successfully defended my PhD dissertation at CVC_UAB! Many thanks to my directors, Joost van de Weijer and Estibaliz Garrote, and to TECNALIA for the support over these years. It was a pleasure to pursue this dream at CVC_UAB and TECNALIA!

With many 🧩 dropping recently, a more complete picture is emerging of LLMs not as a chatbot, but the kernel process of a new Operating System. E.g. today it orchestrates: - Input & Output across modalities (text, audio, vision) - Code interpreter, ability to write & run

Love it.