Keep Pushin'!

Here for both

I used it for a keyword spotting task. Well, it is actually the sota (not by me) as far as I know. I’m using a model trained in such way to explore some particular compression techniques.

Anyway, I’d love to hear more :)

I mean, I think your posts are really useful, so if you can maximize their spread I think everyone would benefit :)

Both!

Hi Michael, I enjoy a lot your photos! Especially the ones with the dog in the nature :)

Yes, you’re definitely right. The fact is that without the proper attitude it becomes overwhelming or you starting jumping around too many things. Nonetheless, I like the positivity you are spreading, tomorrow I will have some time and I will make it worth ;)

The problem is finding the time for all these nice things :( everyday I fall in love with something new

Would love to join

Oh, that’s really nice to hear! I was thinking it was a work meant to be closed source. I’ll wait patiently then

Interesting! I would love to know more but I suppose you already said what you could ;)

Really nice work! Looking forward for the next updates!

Btw, is the model something new or did you “simply” scaled down a well know architecture ?

Honestly my X feed was good enough but many people that I love to interact with moved so I did it

Farfalline rimandate sul nucleare :(

Omg this is so me!

The Ultimate Directory of tools and applications for Bluesky A curated collection of all things relating to the Blue Sky social media platform.

I'm just learning about Starter Packs. Here's a great place to find them! blueskydirectory.com

Welcome!

P.S., I can definetely wrong on some details, not an expert in GPU arch :)

Apple GPU cores should corresponds to the Streaming Multiprocessors (SM) of NVIDIA.
In the Lovelace architecture each should be composed of 128 CUDA cores which are “composed” of a warp of 32 threads…

I think it depends what you want to compare. For raw compute you can compare num of threads

Is there someone working on optimization of ML/DL algorithms? If yes I would love to connect!

Really interesting, thx!

Would love to be added! I did some work with EMG and PPG!

I didn’t think a lot about it, but everything here on bluesky is much more immediate than on mastodon.

In the future I would love to find the time to explore the resiliency of such models to optimization (which is really my field of expertise) such as quantization, pruning, etc. in order to explore the feasibility of deploying them directly on satellite HW

(1/5) Very excited to announce the publication of Bayesian Models of Cognition: Reverse Engineering the Mind. More than a decade in the making, it's a big (600+ pages) beautiful book covering both the basics and recent work: mitpress.mit.edu/978026204941...

I did some (preliminary) work for building damage classification in war scenarios using geospatial VHR images. If u want add me to the list :)

Yes, I think that a light organization may work for me. I’ll try!

It’s beautiful! Where is this place (if you want to say that of course)?

Would love to be added :)

I think I am :(