mrv_asura
mrv_asura on Hacker News
Posted by
mrv_asura on Thursday July 09, 2026 @03:23PM
from the mrdaybird.github.io dept.
Awesome! Curious how you guys manage to make it "75% cheaper"?
Posted by
mrv_asura on Monday January 27, 2025 @09:12PM
from the github.com dept.
Posted by
mrv_asura on Saturday January 25, 2025 @08:51AM
from the ronkathon.pluto.xyz dept.
Posted by
mrv_asura on Saturday January 25, 2025 @08:50AM
from the ronkathon.pluto.xyz dept.
Posted by
mrv_asura on Saturday January 25, 2025 @08:47AM
from the ronkathon.pluto.xyz dept.
Posted by
mrv_asura on Thursday June 20, 2024 @09:04AM
from the github.com dept.
Posted by
mrv_asura on Thursday June 20, 2024 @08:56AM
from the mrdaybird.github.io dept.
Posted by
mrv_asura on Thursday November 09, 2023 @10:27AM
from the archive.org dept.
I learnt about KL Divergence recently and it was pretty cool to know that cross-entropy loss originated from KL Divergence.
But could someone give me the cases where it is preferred to use Mean-squared Error loss vs Cross-entropy loss? Is there any merits or demerits of using either?
Great explanation. Now it makes sense!
I briefly remember napping during my physics lectures looking over the transformation equations, which did not make sense to me. ( I know it's partly my fault but...)
Posted by
mrv_asura on Thursday July 27, 2023 @11:01AM
from the citeseerx.ist.psu.edu dept.
Posted by
mrv_asura on Monday July 24, 2023 @10:49AM
from the modernescpp.com dept.
I have never used Postman, but I am curious, how useful or good, if any, their product is? given the recent bashing it received for being over-valued and being just a wrapper around curl. If people like it and are willing to pay for it, isn't its valuation justified, which I think is true for any startup in general.
Read: Another nail in Netflix's coffin.
Creator of mlpack and armadillo announced Bandicoot.
Posted by
mrv_asura on Wednesday July 19, 2023 @11:01AM
from the coot.sourceforge.io dept.
Question: Does fp16 provide more accuracy than mixed-precision? If so, any reason for this to be happening?
Looking at the discussion, everybody is agreeing to the fact that it is already well-known that fp32 is overkill and fp16(or bf16) is already industry standard(for most cases at least). But any opinions on mixed precision floating point is seems to be missing. Has anybody seen benchmarks that seem to indicate that mixed-precision fp performs worse than fp16 and fp32,(other than the paper)?
> The reason for having more bits is having large numbers of incoming or outgoing connections.
I am having trouble getting my head around this statement, could you please explain this more? This idea is not intuitive to me. Any example will be much appreciated.
My current thought process is this: how having more dynamic range of a single weight/parameter will help in more incoming and outgoing connections?
Maybe I am approaching this statement the wrong way.
Thank you. :)
Well, for me in the beginning getting results matter more than clean code. There are certain things you learn along the way, like what 'not' to do, which help you in the long run, so you don't run into hiccups along the way while doing things quickly.
It's also important to write simple code because simple code is more scalable and refactor-able in general.
I wish I had put more benchmarks in the articles to show its application and how it performs. I had some interesting findings on the performance differences between CPU(unrolled, transposition before matmul etc), GPU and also simd code using WebAssembly.
I hope to put in another article as this blog was already too long.
Posted by
mrv_asura on Thursday May 18, 2023 @02:20AM
from the mrdaybird.github.io dept.
"Financial freedom was never the destination. There is no destination.
There’s only a road.
And if life is a journey, then financial freedom is more like a horse.
Being on a horse makes your journey easier, faster, and more enjoyable."
- a quote from the article.
Posted by
mrv_asura on Wednesday May 10, 2023 @01:44AM
from the coryzue.com dept.
Posted by
mrv_asura on Sunday April 23, 2023 @10:59AM
from the petewarden.com dept.
Posted by
mrv_asura on Friday April 14, 2023 @07:38AM
from the matrixcalculus.org dept.