Offshore
Photo
hardmaru
RT @3wadarch: Stable Diffusion Reimagine
無料で何回かできるよ。
image2imageをやっているようだけど、元画像からの何が写っているかの正確な識別を従来より良い精度でやって、それらのトークンを読み込ませて画像を吐き出しているのだと思う。
https://t.co/LfjtyIGBTS
左上が元の写真やイラスト https://t.co/vZtJ7jxIhn
tweet
Brian Chau (Ottawa 22-24th, DC 29th-2nd)
The world is eating software ...
tweet
Offshore
Photo
Hugging Face
RT @victor_iyi: I remember the early days of @huggingface when I requested that they support TensorFlow. It’s crazy how far they’ve come in just 4 years. Kudos to the 🤗 team. https://t.co/ujinauMDbw
tweet
Yann LeCun
RT @Experiential_AI: COMING SOON: Chief AI Scientist at @Meta @ylecun leads our May 24, 2023 Distinguished Lecturer seminar

Don't wait to reserve your spot in-person at @Northeastern or online. https://t.co/BD7lNz5117

#AIexpert #AIevent #AIBoston #ML #Meta
tweet
MIND SPLITTER
RT @_SullyMusic: ICYMI - I dropped a mix yesterday featuring 15 new ID's and a bunch of unreleased music from my talented friends and you should check it out :)

https://t.co/fSvj7djqrQ
tweet
Offshore
Photo
Hugging Face
RT @younesbelkada: A new release for TRL ! 🤯

1- Larger models support using Naive Pipeline Parallelism: you can now fit 100B scale models and apply RLHF!
2- PEFT + data parallelism: parallelise your training to train on more data and more GPUs

Full release notes here: https://t.co/ZEnTXKzK4O https://t.co/cd7MnfV7YT
tweet
Offshore
Photo
Hugging Face
RT @michelleyhbn: 🤗@HuggingFace Brest office is getting fancy! Love our newest addition 🥰 @guillaume_salou @coyotte508 @dlsjgffskjgfks @scottedelstein https://t.co/xiMCIG4f2Z
tweet
Offshore
Photo
Hugging Face
RT @mrm8488: 44 to 5⃣0⃣0⃣ 🧠 at @huggingface 🤗 Hub! https://t.co/BTa9mMftJi
tweet
Offshore
Photo
Hugging Face
RT @xenovacom: The past 12 hours have been insane! 🤯
🔥 #1 trending on Hacker News
⭐️ 300 -> 700 stars on GitHub
🤗 Shout-outs from the @huggingface team!

I'm so grateful for everyone's support! Stay tuned for more updates!

Code: https://t.co/oWpzo4CQYO https://t.co/xSK9fNiNkL
tweet
Offshore
Photo
Firas Durri
It's funny that the way Facebook Research solved the 'potential bad PR' problem with releasing open source AI models is to only send the model weights upon request to researchers... who then leaked it online https://t.co/hcUXCe3YoM
tweet