Yeah but how do you get to that page? The reason people are confused is the typical Google stuff: there's "AI Studio", but then in Google Cloud there's "Vertex AI". Then there's some other AI product suite as well. To get a paid key I linked up my gcloud project.
@honnibal
-
Competition with Hugging Face: Allies Against Technical Failures
By
–
People have often asked about competition with @huggingface , but actually we have a common enemy: shit not working. Projects not succeeding, or just too difficult to start, is still the big problem in AI/NLP. We welcome all allies in the fight https://
x.com/ClementDelangu
/ClementDelangue/status/1813864232321990992
… -
Returning to hands-on spaCy development after venture funding
By
–
I'm getting back to hands-on @spacy_io development. We raised money in 2021, but building a venture-backed company isn't quite like building a library. I hate how vaguely these things are usually discussed, so I wrote a long post: https://
honnibal.dev/blog/back-to-o
ur-roots
… -
MLOpsWorld Summit Insights and BERT Model Performance Talk
By
–
The @MLOpsWorld Generative AI Summit was great! Thanks all for a super engaging event Lots of interesting conversations, and even more I didn't get to say hi to. Slides from my talk https://
speakerdeck.com/honnibal/how-m
any-labelled-examples-do-you-need-for-a-bert-sized-model-to-beat-gpt-4-on-predictive-tasks
… -
MLOpsWorld Austin: BERT vs GPT4 Talk and Baggage Issues
By
–
Just got to @MLOpsWorld in Austin! Looking forward to my talk tomorrow at 14:00 (“How many labelled examples do you need to beat GPT4 with BERT?”) Unfortunately the airline gate-checked our bags in Berlin and didn’t put them onto our connection to Austin . So we arrived with
-
Training Better TRF Models on Updated Data in Progress
By
–
We're looking to train better trf models on the updated data, yes. The work is still in progress though — plans might vary a bit.
-
spaCy Introduces Premium Subscription Service for Advanced Language Models
By
–
Here's the plan. We want to introduce a subscription service with premium up-to-date models with an easy PyPi server. You'd write: $ pip install spacy-models[pro_en_ner_web_trf,pro_de_core_chat_sm] -f https://$KEY@models.spacy.io …and get models trained on better data!
-
spaCy seeks sustainable funding for development and data
By
–
We need a sustainable way to fund more spaCy development, especially data. If this sounds like something your company could use, sign up & help make sure we're building what you really need. You'll get early beta access & a 20% discount for the 1st year once it goes on sale.
-
spaCy Users Survey: Improving Trained Pipelines and Deployment
By
–
spaCy users! Are you using trained @spacy_io pipelines (en_core_web_sm etc.) at work? We've heard they're doing good work but you want: better accuracy, esp. on new text PyPi server to put them in requirements more vectors & component combos
-
Explanation of backpropagation: max function implements NAND gates
By
–
This explanation was kind of polarising — some people found it pointlessly abstruse, others found it worked for them when others didn't: https://
thinc.ai/docs/backprop1
01
… If you get down to the bit with the two tables, it explains how a max function can implement nand, while a linear