Why everyone talk about _price_for_outsider as a cost? Like "How much would this Guinnes record dinner for 1k person cost if you order it at restaurant". But Google is the restaurant and it costs much less for them to do it.
Wow: Google's "Meena" chatbot was trained on a full TPUv3 pod (2048 TPU cores) for **30 full days** - That's more than $1,400,000 of compute time to train this chatbot model. (! 100+ petaflops of sustained compute !)pic.twitter.com/BiPdTTG5E9
-
-
-
Reproducibility and Replicability in Research?
- Još 4 druga odgovora
Novi razgovor -
-
-
By comparison, Mitsuku takes 40Mb of space, needs 4Mb of RAM and runs on a single CPU. As a bonus, I even let people talk to it!
-
On the plus side, Google's compute is at least 100% renewably-offset. Let's hope if Amazon do this they do it in a green region & not US-East...
- Još 4 druga odgovora
Novi razgovor -
-
-
Any estimate of the kWh ?
-
294,912 kWh
- Još 2 druga odgovora
Novi razgovor -
-
-
Whereas a human capable of the same conversation costs on average only $233,000 to raise.
-
That 18 years to train it, though, not so great
- Još 2 druga odgovora
Novi razgovor -
-
-
The
#Eliza program was able to do the same 55 years ago, and it did not cost 1.4M$. Stop the bullshit, the energy and money waste. We should focus on real research on#AI instead of overselling results. https://en.wikipedia.org/wiki/ELIZA?wprov=sfla1 …pic.twitter.com/vycRgvvVyb
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
Did she mention she's always wanted to go?
-
Oh, nice!
- Još 1 odgovor
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.