ramesh1994

7 ms·

1.
Pg_AI_query – AI-powered SQL generation and query analysis for PostgreSQL(postgresql.org)
2 points by ramesh1994 10mo ago|0 comments
2.
by ramesh1994 2y ago
Sorry I don't know how hackernews notifications work, didn't see this reply. But if you see this here's another reminder for you to clean this up into a public demo :)
3.
by ramesh1994 2y ago
This is really cool! Appreciate sharing this work and the explanation. You mentioned that massaging the data into shape as one of the problems, which I think is possibly one of the best applications of LLMs in my opinion. Creating a pipelin
4.
by ramesh1994 2y ago
Hey this sounds pretty cool! If this is public would you mind sharing a link?
5.
by ramesh1994 3y ago
I just found about this project from this comment, absolutely excited to try this out. As someone who's never used any of the infrastructure tools, I'm thinking of pyinfra as a way to run shell commands + install dependencies on h
6.
by ramesh1994 3y ago
I think distillation in the original sense isn't being done anymore but finetuning on outputs from larger models like GPT-4 is a form of distillation (top-1 logit vs all logits and a curated synthetic data instead of the original datas
7.
Second H-1B lottery announced for FY 2024(uscis.gov)
2 points by ramesh1994 3y ago|0 comments
8.
by ramesh1994 3y ago
Head over to the elueuther.ai discord and discuss with some of the folks there. Tons of small experiments with LLMs can use the $10k in compute
9.
by ramesh1994 3y ago
It prohibits anything that competes with OpenAI services i.e as long as you're not literally providing an LLM API commercially you should be fine
10.
by ramesh1994 3y ago
> This means it is way cheaper to look something up in a vector store than to ask an LLM to generate it. E.g. “What is the capital of Delaware?” when looked up in an neural information retrieval system costs about 5x4 less than if you as
11.
by ramesh1994 3y ago
I think parts of the write-up are great. There are some unique assumptions being made in parts of the gist > 10: Cost Ratio of OpenAI embedding to Self-Hosted embedding > 1: Cost Ratio of Self-Hosted base vs fine-tuned model queries I
12.
Paper Summary:TinyViT Fast Pretraining Distillation(efficientml.substack.com)
1 points by ramesh1994 3y ago|0 comments
13.
Paper Summary: EL-Attention: Memory Efficient Lossless Attention for Generation(efficientml.substack.com)
1 points by ramesh1994 3y ago|0 comments
14.
by ramesh1994 3y ago
I've been looking for a course like this! Especially great given how much of the recent progress in training large models is made possible with the aid of flash attention and fused kernels
15.
by ramesh1994 3y ago
Was it for seeding/hosting torrents or from just downloading them? How long did the whole thing take to play out? I've always assumed that consuming torrents has been low stakes to the point where its not worth any enforcement
16.
by ramesh1994 3y ago
It is a pretty fun game https://wiki-race.com/
17.
by ramesh1994 3y ago
The term "chinchilla" predates llama/alpaca. It doesn't directly map to a specific model, rather a family of compute-optimal models.
18.
by ramesh1994 3y ago
That fact that OpenAI decided to not even reveal the parameter counts / training data composition in their "technical" report is surely a sign that their moat isn't as big as it would seem in fear of competition. This pa
19.
Emergent Deception and Emergent Optimization(bounded-regret.ghost.io)
2 points by ramesh1994 4y ago|0 comments
20.
The Crypto Elites Are Plotting a Wall Street Merger(concoda.substack.com)
32 points by ramesh1994 4y ago|23 comments
21.
by ramesh1994 5y ago
I would also highly recommend the blog post series [1] from Cliqz talking about the tech behind the search. [1] - https://0x65.dev/
22.
by ramesh1994 5y ago
I think it is definitely a hard problem to solve on a large scale to address latency, quality and size of the index they plan to address. It definitely isn't as easy as spinning up an elastic search cluster. I agree that getting someth
23.
Microsoft DeepSpeed: Zero-Infinity(microsoft.com)
6 points by ramesh1994 5y ago|0 comments
24.
by ramesh1994 7y ago
I've taken Georgia Tech's OMSCS program. What really sets this apart is the active feedback on assignments / course interaction from Professors and TAs in Piazza. Current online courses have really watered down content and ze