Sebastian Raschka
speaker
1,024 appearances
1 recordings
1 series
first heard Feb 2026
last heard 1 Feb
Sebastian Raschka’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsRecordings per month over the last 12 months — 1 in all, peaking in Feb 2026 with 1.
Appearances
Stable Diffusion was a company.
Other companies built their own diffusion models.
And then people are now like, okay, can we try this also for text?
Doesn't make intuitive sense yet because it feels like, okay, it's not something continuous like a pixel that we can differentiate.
It's like a discrete text.
So how do we implement that denoising process?
But
It's kind of like similar to the BERT models by Google.
Like when you go back to the original transformer, so they were like the encoder and the decoder.
The decoder is what we are using right now in GPT and so forth.
The encoder, it's more like a parallel, let's say, technique where you have multiple tokens that you fill in in parallel.
So GPT models, they do autoregressive one token at a time.
You complete the sentence one token at a time.
And in BERT models, you have a text, let's say a sentence that has gaps, you mask them out.
And then one iteration is filling in these gaps.
And text diffusion is kind of like that, where you are starting with, let's say, some random text, and then you are filling in the missing parts, or you're refining them iteratively, and you have multiple iterations.
And the cool thing here is that this can do multiple tokens at the same time.
So it's kind of like
the promise of having it more efficient.
Now, the trade-off is, of course, well, how good is the quality?
Showing 641–660 of 1,024 · page 33 of 52
← Previous
Next →