What about using @livebookdev to spawn 64 machines with GPU on @flydotio, each machine fine-tuning BERT with different hyperparameters and graphing in realtime in 2-3 minutes? @cigrainger shows how at @ElixirConf!
I see folks minimising the value of OTP and websockets with Phoenix and LiveView. Lots of 'oh cool but it's only good for chat or social media'. A quick 🧵. #myelixirstatus
My #ElixirConfEU keynote is up! I'm hoping it helps folks gain the confidence to put Nx into production. Machine learning in Elixir is production ready _today_. #myelixirstatus
@jjen_abel This pretty accurately captures our experience. Wish it hadn’t taken years to learn all of these lessons. AI is hot now so it may help start a conversation, but it can also make it harder to close.
Fantastic keynote from Chris Grainger about one year with Machine Learning in Elixir at CodeBEAM EU.
With LiveView and Nx, they went from three teams (front, backend, data) to a single Elixir team, which is extremely valuable for startups (which need to move fast!)
@levie Yeah it’s spooky. We do this with patents. You can summarize a US patent in Thai. Or a Japanese one in Spanish. Huge for democratizing knowledge and scientific progress.
Join to hear the industries top IP expert sharing their Exclusive Presentation on "Patent search as data curation: knowledge work in the age of AI" and connect with them just at The IPR Gorilla - Virtual IP Conference on 23rd February, 2023 GMT Zone. Register today only:
Not even 2 years ago we were struggling to get autograd right so we could train MNIST, and now we have transformers in Elixir!
In 2023 my goal is to be able to share success stories of Nx, Axon, and Bumblebee in production at scale! And hopefully I can share one pretty soon ;)
I asked #Galactica about some things I know about and I'm troubled. In all cases, it was wrong or biased but sounded right and authoritative. I think it's dangerous. Here are a few of my experiments and my analysis of my concerns. (1/9)