rom1504

8 ms·

1.
by rom1504 2mo ago
Why is a very good question indeed! My guess is we have not figured out how to include long term enough tasks into the RL training. As to whether they can today, it's pretty easy to figure out that they can't by observing the diff
2.
by rom1504 2mo ago
I think the simple answer is LLM cannot do long term planning. Maintainability means thinking what will happen to this code in the next year and possibly hundreds of changes and based on that selecting the right abstractions that will work
3.
Imagine it, create it: Veo 2 is coming to YouTube Shorts(blog.youtube)
1 points by rom1504 2y ago|0 comments
4.
by rom1504 3y ago
yeah it's sad
5.
by rom1504 4y ago
Hehe, well you know, PR welcome, the front end is 500 lines https://github.com/rom1504/clip-retrieval/blob/main/front/sr... Other people have done a few alternate front ends already This one is mean
6.
by rom1504 4y ago
yes indeed. Video is the clear next step.
7.
by rom1504 4y ago
Looks like you missed the whole point of this dataset. The idea that we proved is you can get a dataset with decent caption and images (that do match yes, you can see for yourself at https://rom1504.github.io/clip-retrieval&
8.
by rom1504 4y ago
Done https://github.com/rom1504/clip-retrieval/commit/53e3383f58b... Using clip for searching is better than direct text indexing for a variety of reasons but here for example because it matches better what s
9.
by rom1504 4y ago
This is due to the aesthetic scoring in the UI. Simply disable it if you want precise results rather than aesthetic ones. It works for your example I guess I'll disable it by default since it seems to confuse people
10.
by rom1504 4y ago
Hi, laion5b author here, Nice tool! You can also explore the dataset there https://rom1504.github.io/clip-retrieval/ Thanks to approximate knn, it's possible to query and explore that 5B datasets with only 2TB of
11.
by rom1504 4y ago
Yeah we have regular contact with many EAI members. EleutherAI is pretty great at open research.
12.
by rom1504 4y ago
Let me start by saying that laion is a non profit, open to anyone that want to contribute. Agreed about the website css. Do you want to contribute? What's the problem with the dataset name exactly? Seems to work pretty well. Yes the da
13.
by rom1504 5y ago
a safe mode is now in place
14.
by rom1504 5y ago
That's interesting indeed! Note that the description search as well as the image search are both using a knn index on embeddings and not exact search. That helps for finding semantically close by items but indeed for exact reference ma
15.
Prismarine-web-client: A Minecraft client in the browser(webclient.prismarine.js.org)
3 points by rom1504 6y ago|0 comments
16.
Semantic search with embeddings: index anything(rom1504.medium.com)
1 points by rom1504 6y ago|0 comments
17.
DeepR – Training TensorFlow Models for Production(medium.com)
1 points by rom1504 6y ago|0 comments
18.
Embeddings are amazing Want to build visual search using any image dataset?(medium.com)
1 points by rom1504 6y ago|0 comments
19.
by rom1504 8y ago
I wrote an overview of what methods and softwares are available for computer vision. Computer vision is really powerful these days !
20.
Overview of methods and softwares for computer vision(medium.com)
2 points by rom1504 8y ago|1 comments
21.
by rom1504 11y ago
For example the chakra engine has a better coverage of es6 than v8. see edge score there https://kangax.github.io/compat-table/es6/
22.
by rom1504 11y ago
So I tried it and it's 72135 bytes using tar czf (created by `fs.writeFileSync('/tmp/file',new Buffer(74207281).fill("1"),'utf8')`) Not that it's the most efficient representation at all, 74