danielcampos93

8 ms·

31.
Evaluating Text to SQL Models Without Execution Accuracy(medium.com)
2 points by danielcampos93 3y ago|0 comments
32.
Snowflake Launches GenAI tools and copilot(techcrunch.com)
1 points by danielcampos93 3y ago|0 comments
33.
by danielcampos93 3y ago
Great paper! I have a question on how much this can be attributed to portions of text which GPT has already seen. Like if the same approach was followed on a new sliver of Wikipedia which GPT had not seen would the results be the same?
34.
REKA releases Yasa, multimodal foundational model(reka.ai)
3 points by danielcampos93 3y ago|2 comments
35.
by danielcampos93 3y ago
This feels like the corporate regurgitation of character.ai. I'm pretty curious what the contracts/paydays look like for the licensed personalities.
36.
by danielcampos93 3y ago
Holy smokes thats a lot of authors
37.
by danielcampos93 3y ago
What happens to Cohere with this announcement? Seems the music has stopped and there is no hyperscaler left. Oracle?
38.
by danielcampos93 3y ago
Most other search engines train with a target of google or with some form of reward which is bootstrapped on google rankings. It makes Bing results implicitly have the same behavior as Google. DDG and others just use BingAPI so googles ince
39.
AI Grant 2 Batch Announced(aigrant.com)
2 points by danielcampos93 3y ago|0 comments
40.
by danielcampos93 3y ago
See. When you have to convince corporate lawyers and security folks that whole switching licenses makes them uneasy. BD and legal are much happier to deal with Meta than the UAE.
41.
by danielcampos93 3y ago
Last week I got an email from them about they had only 6000 H100 left. Got one today about how they have 4000. Tracks
42.
by danielcampos93 3y ago
nice of them to wait for the end of final
43.
by danielcampos93 3y ago
Thats because they have an army of sell-swords via MSFT
44.
Neeva acquired by Snowflake(snowflake.com)
166 points by danielcampos93 3y ago|111 comments
45.
Neeva launches its generative AI search engine internationally(techcrunch.com)
3 points by danielcampos93 4y ago|0 comments
46.
by danielcampos93 4y ago
Using this shows me how bad my handwritting is. Have to turn it down to minimal legibility to make it look like mine.
47.
by danielcampos93 4y ago
Common practice. In places where I have worked if you told the company you had a competitive offer the standard practice was to take 15 minutes to try to get you a matched offer. If you declined or they couldn’t match it was automatic walk
48.
Apple Vehicle announcements coming in late 2023
1 points by danielcampos93 4y ago|1 comments
49.
by danielcampos93 5y ago
Maybe not direct inference but for generation of dense represenatations that can be used with their nearest neighbor search system this large cluster will probably be hyper userful.
50.
by danielcampos93 5y ago
I have had little luck getting them on AWS. If you time it outside of some major paper deadline I have had really good success on GCP. If you are willing to break out of the big clouds CoreWeave has great ability and even better prices. Hig
51.
Meta scales Language Model to 1.1T parameters using Mixture of Experts(twitter.com)
2 points by danielcampos93 5y ago|0 comments
52.
by danielcampos93 5y ago
This move was a defensive move. Keep any other MEGA-CORP from getting access.
53.
by danielcampos93 5y ago
Neural Magic | Machine Learning Engineer, Full Stack Engineer, Product Manager, ML Ops Engineer | Somerville, MA | https://www.neuralmagic.com Neural Magic is a software company challenging the notion that specialized hardware l
54.
Facebook Is Now Meta(twitter.com)
3 points by danielcampos93 5y ago|1 comments
55.
by danielcampos93 5y ago
T5 has a notion of prompting. None of the *BERT have notions of prompting.
56.
Hugging Face Announces Infinity:Ultra-Fast Inference in Your Own Infrastructure(huggingface.co)
2 points by danielcampos93 5y ago|0 comments
57.
Faster and smaller Hugging Face BERT on CPU via compound sparsification(reddit.com)
2 points by danielcampos93 5y ago|1 comments
58.
by danielcampos93 5y ago
combining multiple sparsification methods to improve the Hugging Face BERT base model (uncased) performance on CPUs. We combine distillation with both unstructured pruning and structured layer dropping. This “compound sparsification” techni
59.
by danielcampos93 6y ago
As much as he might stretch the truth, the most beautiful Elon trait is he dreams and is not afraid to share this dream world in the world no matter how foolish, foolhardy, or stupid it is. The rest of the FAANG CEOs provide a vague stateme
60.
by danielcampos93 6y ago
I believe the Bing API can do this. "Use smart identification of image content to recognize celebrities, find products, or search for related content." https://azure.microsoft.com/en-us/services/cognitive

More ›