lnyan

lnyan

8 bulan yang laludiscuss

Can it detect when it itself is down

lnyan

tahun laludiscuss

normalizing flow might be unpopular but definitely not a forgotten technique

lnyan

tahun laludiscuss

Note that it's very likely that Veo models are based on "Flow Matching" [1]

[1] https://arxiv.org/abs/2210.02747

lnyan

tahun laludiscuss

I just pre-ordered a Bigme 13-inch color e-ink monitor with a 30Hz refresh rate. At $499, it's much cheaper than the Dasung, though the quality and e-ink display algorithms might not be as good

lnyan

2 tahun yang laludiscuss

I've been using a BOOX Tab Ultra C Pro (with the Kaleido 3 color e-ink) for a year. The colors are quite dim, to the point where I sometimes forget it's a color e-ink display. The Colorsoft display looks significantly better in comparison (even seems to have better contrast than ClearInk?), so I'm curious how it works.

lnyan

2 tahun yang laludiscuss

Slightly off-topic: Is there a library in JAX that supports post-training quantization, similar to the one mentioned?

lnyan

2 tahun yang laludiscuss

The post was moderated within 10 minutes of being submitted: https://news.social-protocols.org/stats?id=40710417

lnyan

2 tahun yang laludiscuss

pure css without background-image also works: https://css-shape.com/film-strip/

lnyan

2 tahun yang laludiscuss

Seems that pasteurization cannot completely inactivate H5N1

> In addition to the mice studies, the researchers also tested to determine which temperatures and time intervals inactivate H5N1 virus in raw milk from dairy cows. Four milk samples with confirmed high H5N1 levels were tested at 63 degrees Celsius (145.4 degrees Fahrenheit) for 5, 10, 20 and 30 minutes, or at 72 degrees Celsius (161.6 degrees Fahrenheit) for 5, 10, 15, 20 and/or 30 seconds. Each of the time intervals at 63℃ successfully killed the virus. At 72℃, virus levels were diminished but not completely inactivated after 15 and 20 seconds

lnyan

2 tahun yang laludiscuss

`import jax.numpy as np`, then we also get a jax implemention after certain modifications: e.g. remove in-place index assignment, replace unsupported functions, etc

lnyan

2 tahun yang laludiscuss

You can use mailchannels to send emails with cloudflare workers

https://blog.cloudflare.com/sending-email-from-workers-with-...

lnyan

2 tahun yang laludiscuss

no, I think it might be

- LLM(prompt_0) = arch/spec of LLM

- LLM(prompt_1) = full weights of LLM

Note that it does not conform the definition of quine as a quine takes no input.

Anyways, constructing a transformer that can autoregressively output its weights would be quite interesting.

lnyan

2 tahun yang laludiscuss

Just wondering if it's possible to achieve LLM quine.

lnyan

2 tahun yang laludiscuss

Will there be Gemma-vision models or multimodal Gemma models?

lnyan

3 tahun yang laludiscuss

This is bad news for me. I am working on a simliar project (gaussian splatting + dynamic scene). Our method is different with the mentioned 4D gaussian splatting, but I am unsure shall I continue or not.

lnyan

3 tahun yang laludiscuss

For the country one, I got "www.google.com/maps/embed?pb=!4v1687119352692!6m8!1m7!1sINHBz4HdSwMAAAQrBnftjg!2m2!1d9.080961517214682!2d7.524398838108427!3f84.34!4f-4.950000000000003!5f0.4000000000000002"

I think it should be 9.08N, 7.52E and it works

lnyan

3 tahun yang laludiscuss

https://github.com/koreader/koreader

koreader has a pdf reflow mode

lnyan

4 tahun yang laludiscuss

Previous post: https://news.ycombinator.com/item?id=25244872

lnyan

4 tahun yang laludiscuss

Also check out Alyssa Rosenzweig's "Dissecting the Apple M1 GPU" series

- https://rosenzweig.io/blog/asahi-gpu-part-1.html, https://rosenzweig.io/blog/asahi-gpu-part-2.html, https://rosenzweig.io/blog/asahi-gpu-part-3.html

- https://rosenzweig.io/blog/asahi-gpu-part-4.html, https://rosenzweig.io/blog/asahi-gpu-part-5.html, https://rosenzweig.io/blog/asahi-gpu-part-6.html

lnyan

4 tahun yang laludiscuss

Seems that HN automatically replaces "MiSTer" with "Mister" when I submitted this post