4 ms·
How many parameters? The LaMDA paper describes 2B, 8B, and 137B parameter variations. Sundar says "We're releasing initially with our lightweight model version
by dzdt 4y ago
How many parameters?
The LaMDA paper describes 2B, 8B, and 137B parameter variations. Sundar says "We're releasing initially with our lightweight model version of LaMDA. This much smaller model..."
So it sounds like the 2B or 8B model variations, compared to GPT3's 175B. [edit: corrected]
I can't imagine such a smaller model coming across as anything near as impressive as ChatGPT.
So the announcement today is for a much smaller, more limited thing than ChatGPT and even that is not actually available today nor even with an announced release date.
- dougmwne 4y agoYeah, I think you’re onto something here. Clearly we won’t know till we can use them side by side, but there’s lots of hints that Google is a ways off from matching ChatGPT. A small model, a closed beta, a press release annd mock-up instead of a link, chest pounding about how they invented it all anyway.
- cypress66 4y agoGPT3 is 175B parameters, but I agree with you.
- dzdt 4y agoOops. Corrected. Thanks.
- 10xDev 4y agoTo compare models strictly on parameters is simply naive. And if we are to compare them only on such basis, smaller models have often proven to be more stable and efficient.