Viewing profile — Azantys

Dysfunctional vibeslop

Here you have shown yourself that progress slows down and doesnt speed up. 8.9/0.35 = ~25x more performance in 10 years from 2006 to 2016. 104.8/8.9 = ~12x more performance in 10 y…

You already can with AMD/Intel gpus. Kraid is a compiler for Arm Mali which is mostly used on mobile stuff, nothing AI related.

I questioned the speed not the output quality, that is another discussion

? Readme says 60-70s per token

0.01 tk/s is unusable for anything, you would wait a whole day for just 1000 token of output, what is the point of projects like this?

Free month is insane for a bit of downtime

Just looking with ncdu/du or another tool which sorts folders by size would have been much faster and easier, was there any need to use Claude here?

If you could render fully in OpenCL and output it via another device it would probably work. But they dont have ROPS so they cant do traditional video rendering. Even something lik…

It really depends on the language, popular languages work pretty good

You are comparing a 35B models to a 635B+ frontier model, of course thats not even close

Old datacenter GPUs could game, but new ones can only do OpenCL/CUDA/ROCm etc. and have no display out. Im using an MI50 right now and I would wish newer datacenter cards could als…

Useless comment

The wording wasnt very good I ment compared to programming or math the amount of logic and reasoning is small (Research level math hardly compares to writing a book in raw reasonin…

Of course there is logic but its nowhere near the complexity of math or programming

Do people really use 100B+ models for writing? I am no writer but to me it seems like writing is one of the easiest tasks with barely any logic or reasoning and as long as its not …

I think model training is pretty hard to do efficiently on a vastly distributed network. If the model cant fit into the VRAM of the node your performance becomes so bad its useless…

+1

Game studios should choose CryEngine/Decima again over UE5

Just dont upgrade the Mainboard firmware then

Yeah but it wasnt close to Opus etc. Still a good local model when it released

Galaxy AI 3.8-Flash-Plus Max (xhigh)

Isnt that more Perplexitys thing anyways?

Career and personal advice from LLMs, not sure if thats your best bet

Is LPDDR5X not too slow for inference, atleast compared against HBM?