Dowwie

19 ms·

241.
by Dowwie 3y ago
You want to see the source code of the transformer that created the compiled llm model, the dictionary corresponding to token values, etc. Instead, you got the weights and binary. I don't know what's going on in this discussion t
242.
by Dowwie 3y ago
Any thoughts here on Fly's Corrosion? https://github.com/superfly/corrosion
243.
by Dowwie 3y ago
What is your use case for DuckDB?
244.
by Dowwie 3y ago
soon, we'll be voice-interacting with an AI assistant about images taken from microscope slides
245.
by Dowwie 3y ago
This article isn't covering the lack of structured concurrency and the blocking dependency on async Drop. The resulting state of async Rust includes leaks and inadequate task management. This isn't a Tokio problem but a Rust pro
246.
Deep learning enabled fast 3D brain MRI at 0.055 Tesla(science.org)
1 points by Dowwie 3y ago|0 comments
247.
by Dowwie 3y ago
If there's a person who studies the works of George R.R. Martin, spending years cultivating writing skills based on just his works, and then writes new material, can that person lose a lawsuit for being too similar in style?
248.
by Dowwie 3y ago
this photo from the article amazes me: https://clan.cloudflare.steamstatic.com/images//3703047/a37c... that's a packed stadium, up to the nose bleeds
249.
by Dowwie 3y ago
An optometrist's office, in New Jersey, presented two pricing options to me, based on vision insurance benefits, which I explain far more clearly than they did with me: pay more out of pocket for just the eye exam and contact lens fit
250.
by Dowwie 3y ago
Roger Murdoch has entered the chat
251.
by Dowwie 3y ago
If I paid $20 to fine-tune a model to do X, and you paid $20 to fine-tune a model to do Y, is there a way to merge models, aggregating X and Y training, without training from scratch again?
252.
by Dowwie 3y ago
How heavy of a lift is it to optimize inference?
253.
by Dowwie 3y ago
Actix-web uses the single threaded Tokio runtime per physical core. This architecture is harder to design for than multi threaded async Tokio. Performance gains aren't worth the effort.
254.
A study of malicious code in PyPI ecosystem(about.honywen.com)
63 points by Dowwie 3y ago|37 comments
255.
by Dowwie 3y ago
myvar = 5 myvar = if somepredicate() do 200 else myvar end
256.
What Makes Good In-Context Demonstrations for Code Intelligence Tasks with LLMs?(arxiv.org)
1 points by Dowwie 3y ago|0 comments
257.
An Empirical Study of Malicious Code in PyPI Ecosystem(about.honywen.com)
3 points by Dowwie 3y ago|0 comments
258.
The Next Frontier in Software Development: AI-Augmented Software Development(computer.org)
1 points by Dowwie 3y ago|0 comments
259.
by Dowwie 3y ago
'jq' is a lot easier to use with ChatGPT assistance. Simply provide sample JSON and tell the LLM how you want to use JQ to parse and transform.
260.
Ask HN: Custom Instructions for ChatGPT Pro?
1 points by Dowwie 3y ago|0 comments
261.
by Dowwie 3y ago
It looks like Falcon is less efficient than Llama, requiring more than double the inputs to yield a barely-better score. Would a Llama-180B outperform?
262.
by Dowwie 3y ago
This economy seems to consist largely of cheaply made products with high profit margins. You're attacking profit margins, so good luck with that. Sorry for the pessimism but this country just derailed the FTC attempt to abolish non-c
263.
by Dowwie 3y ago
Given how expensive it is to train, my impression is that the world in 2023 generally cannot afford to experiment with custom trained models and only well-funded organizations can within a range of acceptability. The risk of spending $20MM
264.
by Dowwie 3y ago
I can create delicious Korean recipes because of Maangchi. I can also create the perfect crispy breakfast potatoes because of J. Kenji Lopez Alt. Cooking videos gave me the confidence to try to replicate what is made. I never feel the sa
265.
Rare Super Blue Moon rises tonight(space.com)
11 points by Dowwie 3y ago|1 comments
266.
by Dowwie 3y ago
Responsive languages with tight feedback loops are best for the analysis of data. Python is very good at this, as is Elixir. Both languages are extensible, featuring foreign function interfaces allowing applications to communicate with fa
267.
by Dowwie 3y ago
It wasn't too long ago that engineers had to walk over hot coals to get their company to use Python. All of the things you're saying were applicable to Python when the company "only" had programmers who knew Java.
268.
by Dowwie 3y ago
the title is literally the first sentence in the article
269.
SEC charges Impact Theory for unregistered offering of NFTs(sec.gov)
216 points by Dowwie 3y ago|244 comments
270.
by Dowwie 3y ago
Rage blogging is never by contributors

More ›