3 ms·
DALL-E 3 uses T5-XXL for the text encoder not GPT-4. The huge leap comes mostly from using text descriptions that aren't trash. https://cdn.openai.com/papers/d
by famouswaffles 3y ago
DALL-E 3 uses T5-XXL for the text encoder not GPT-4. The huge leap comes mostly from using text descriptions that aren't trash.
https://cdn.openai.com/papers/dall-e-3.pdf https://cdn.openai.com/papers/dall-e-3.pdf
- Sharlin 3y agoHmm, that paper is interesting, and definitely descriptions are a big bottleneck in training. But they only say they used T5-XXL in the experiments related to the paper. Who knows what Bing is actually running?