Firstname Lastname
takeraparterer
AI & ML interests
None yet
Recent Activity
replied to
TuringsSolutions's
post
about 10 hours ago
I created something called 'Hyperbolic Embeddings'. I literally just embed the tokens into Hyperbolic Space instead of Euclidean space. At first, this did not get me the gains I was expecting. I was a sad panda. Then I thought about it, a Hyperbolic Embedding needs a Hyperbolic Optimizer. So, instead of Adam, I used Riemannian Adam (RAdam). "Ladies and Gentlemen, We Got 'Em!"
replied to
TuringsSolutions's
post
about 11 hours ago
I created something called 'Hyperbolic Embeddings'. I literally just embed the tokens into Hyperbolic Space instead of Euclidean space. At first, this did not get me the gains I was expecting. I was a sad panda. Then I thought about it, a Hyperbolic Embedding needs a Hyperbolic Optimizer. So, instead of Adam, I used Riemannian Adam (RAdam). "Ladies and Gentlemen, We Got 'Em!"
replied to
TuringsSolutions's
post
about 11 hours ago
I created something called 'Hyperbolic Embeddings'. I literally just embed the tokens into Hyperbolic Space instead of Euclidean space. At first, this did not get me the gains I was expecting. I was a sad panda. Then I thought about it, a Hyperbolic Embedding needs a Hyperbolic Optimizer. So, instead of Adam, I used Riemannian Adam (RAdam). "Ladies and Gentlemen, We Got 'Em!"
Organizations
takeraparterer's activity
replied to
TuringsSolutions's
post
about 10 hours ago
replied to
TuringsSolutions's
post
about 11 hours ago
*fuck
replied to
TuringsSolutions's
post
about 11 hours ago
bro im sgd optimizer
replied to
TuringsSolutions's
post
about 11 hours ago
I don't think I'm who you think I am
replied to
TuringsSolutions's
post
about 13 hours ago
Ok
replied to
TuringsSolutions's
post
about 13 hours ago
Whos cr
replied to
TuringsSolutions's
post
about 14 hours ago
I think you should add dropout or decrease the size of the model
replied to
TuringsSolutions's
post
about 14 hours ago
I think your model is overfitting you should add dropout or decrease the size of it
replied to
TuringsSolutions's
post
about 14 hours ago
Why so serious?
replied to
TuringsSolutions's
post
about 15 hours ago
You should try dropout or decreasing the model size
replied to
TuringsSolutions's
post
1 day ago
What about test loss? It looks like overfitting to me.
replied to
TuringsSolutions's
post
1 day ago
What about test loss?
replied to
TuringsSolutions's
post
6 days ago
why so serious?
replied to
their
post
7 days ago
glad you like it!
replied to
TuringsSolutions's
post
7 days ago
that's a FFN, which is only a small part of an LLM
Reacted to
TuringsSolutions's
post with 😔
7 days ago
Post
448
What if I told you that LLM models do not simply predict the next token in a sequence but instead utilize an emergent structural pattern-based system to comprehend language and concepts? I created a graph-based optimizer that not only works, but it also actually beats Adam, like very badly. I prove it thoroughly using SMOL LLM models. The secret? The graph is not what you think it is, humans. Code, full explanation, and more in this video. The Rhizome Optimizer is MIT licensed. I have completed my research. I fully understand now.
https://youtu.be/OMCRRueMhdI
https://youtu.be/OMCRRueMhdI
-
- 6 replies
replied to
TuringsSolutions's
post
16 days ago
has threatened violence in subtle ways
😭😭😭😭😭😭
replied to
TuringsSolutions's
post
16 days ago
very intriguing. looking into this 👀👀👀
replied to
TuringsSolutions's
post
16 days ago
Reported.
tell me more
Reacted to
TuringsSolutions's
post with 😔
16 days ago
Post
3944
Are you familiar with the difference between discrete learning and predictive learning? This distinction is exactly why LLM models are not designed to perform and execute function calls, they are not the right shape for it. LLM models are prediction machines. Function calling requires discrete learning machines. Fortunately, you can easily couple an LLM model with a discrete learning algorithm. It is beyond easy to do, you simply need to know the math to do it. Want to dive deeper into this subject? Check out this video.
https://youtu.be/wBRem2p8iPM
https://youtu.be/wBRem2p8iPM
-
-
- 8 replies