4 ms·
Mainly because global video data corpus is > 100k larger than global text corpus, so you will need to train much larger models for much longer (than current LLM
by kadushka 1y ago
Mainly because global video data corpus is > 100k larger than global text corpus, so you will need to train much larger models for much longer (than current LLMs).