Galactica: an AI trained on humanity's scientific knowledge (by Meta)

Re:Galactica: an AI trained on humanity's scientific knowled... by mark_l_watson on Tuesday November 15, 2022 @10:26PM (#33618444)

What is the easiest way to try this via an API? I like that after reading about Flan-F5, that I could run their large model (not the extra, extra large) on Colab, loading the model from Hugging Face. I have been using GPT-3 for about a year via OpenAI.

In a way this was an obvious idea, but still major kudos for Meta for doing it. Hopefully there will be an affordable API for using it.

[ Parent | on HN ]

Re:Galactica: an AI trained on humanity's scientific knowled... by touringa on Wednesday November 16, 2022 @12:25AM (#33619264)

https://github.com/paperswithcode/galai

[ Parent | on HN ]

Re:Galactica: an AI trained on humanity's scientific knowled... by mark_l_watson on Wednesday November 16, 2022 @02:25PM (#33629042)

Thanks so much!

[ Parent | on HN ]

Re:Galactica: an AI trained on humanity's scientific knowled... by nealmcb2 on Saturday November 26, 2022 @01:12PM (#33755088)

You can run Galactica (the "base" model) for free on Colab (https://colab.research.google.com/). It takes about 4 minutes to start up. Just specify a "GPU" runtime on colab, and follow the simple instructions from their github (https://github.com/paperswithcode/galai):

!pip install galai

import galai as gal

model = gal.load_model("base")

model.generate("Scaled dot product attention:\n\n\\[")

[ Parent | on HN ]