Eduard (Edward) Dulharu
speaker
420 appearances
1 recordings
1 series
first heard Aug 2026
last heard 7 Aug
Eduard (Edward) Dulharu’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 Aug 2026 with 1.
Appearances
to see what is the loss, but the ultimate test is to uh empirically see if the model um
Oh
really understood the the patterns and did not memorize the answers.
So in I uh this was one of my um
Um
One of my challenge to uh to to solve this memorization problem and I solve it by okay.
Okay.
Yes, we can go there.
Yeah.
And this is this was one uh challenge I need to uh f I I faced and uh I solved it, I think, uh by um first uh teaching a model uh fundamentals like C CNA, what is subneting and um even though the models are not native in math, they are not.
Uh I I started with
uh with fundamentals with uh fundamental consumps with the networking.
But but the most
It means that it should see a lot of uh say subnetting examples in order to understand why we using uh pattern recognitions what uh what is a subnet and what is un We can't teach it how to subnet, but you can show it enough subnetting examples that it should be able to give us the correct answer.
As we do, as we do.
No, it doesn't.
Yeah.
Okay.
Yeah.
So the objective of his uh um first level, which is supervised fine-tuning, is to um um develop an understanding of the model to be able to generalize.
Showing 261–280 of 420 · page 14 of 21
← Previous
Next →