Henok | Neural Nets
1.61K subscribers
233 photos
20 videos
13 files
157 links
Download Telegram
How is everyone an expert in AI these days😳?

Just because you used openai's api, you don't just become an "AI Expert" or "Gen AI Expert"(idk what this even mean) or LLM expert(I'm not sure if these people even understand LMs).

The amount of hype you see on LinkedIn is because of such people.
🤣15👍1
If you like Computer Vision, make sure to join this next week on October 1. I will be hosting it at C4AI.
🔥13
OpenAi is nothing without Alec Radford.

The role of Alec Redford, first author of some of the coolest things OpenAI did like GPT1, GPT2, and Whisper, coauthor on DallE, GPT3, PPO, Scaling laws, RLHF, is not sufficiently appreciated.

Hired into OpenAI as a 23 year old freshly out of undergrad (aka No PhD, just a small startup), seems to have led OpenAI's experimentation with transformers.
🔥144🥰1
Forwarded from Theodore Discovery ∆
Every holiday i receive a happy holiday wish via SMS from banks I don't remember signing up. TBH I don't even know if some exist.

When did I sign up to ፀሐይ ባንክ, አንበሳ ባንክ.

Any way እንኳን አደረሳችሁ for those of you who are subscribed to this channel.
😁10
Forwarded from Techኢት (Alpha)
🚨🚨Reminder🚨🚨


Addis Ababa University's Vice President for Institutional Development, Dr. Wondwossen Mulugeta, will be joining us on next week's Techኢት Podcast! 🎙

Tune in on October 1, 2024, at 8:00 PM (2:00 LT) for insights on AI, natural language processing (NLP) for Amharic, and tech innovation at Addis Ababa University and beyond. Don’t miss out! 🎧

@techinethio
🔥7
Techኢት
🚨🚨Reminder🚨🚨 Addis Ababa University's Vice President for Institutional Development, Dr. Wondwossen Mulugeta, will be joining us on next week's Techኢት Podcast! 🎙 Tune in on October 1, 2024, at 8:00 PM (2:00 LT) for insights on AI, natural language processing…
He was one of the first 2 people I met back in high school doing NLP for Ethiopian languages, which I'm grateful for. I'm sure you will learn a lot and be inspired to explore the field.
13
Forwarded from Beka (Beka)
Introducing Better Auth

The most comprehensive authentication library for Typescript.

https://better-auth.com
https://github.com/better-auth/better-auth
🔥72
Resources to study RL:

- Introduction to RL book by barto and sutton

https://spinningup.openai.com/en/latest/ great resource includes papers, RL resources, concepts, RL frameworks

- Hugging face deep RL course: https://huggingface.co/learn/deep-rl-course/en/unit0/introduction


Source
🔥6👍4
Addiscoder was one the best things I did in high school 🔥. It was truly a life changing experience. I also went back as a TA and it was even more awesome.

If you are in university (or anyone interested in it), good in data structures & algorithms, please be a TA and be a member of the amazing family.

If any high schooler here, you can also apply to be a student.

More info here

https://www.addiscoder.com/
9
Question 💡

What is enthusiast in a LinkedIn profile mean? Like AI Enthusiast, Data science enthusiast etc. BTW I'm not saying it's a good or bad thing.

I'm genuinely curious about this.
😁4
My first hacktoberfest was all in for the tee lol. It was nice to have it but it was an enjoyable thing regardless of it.
🔥101
Forwarded from Techኢት (Alpha)
Techኢት Podcast S02E14 is out! 🎙

We had an amazing conversation with Dr. Wondwossen Mulugeta , VP for Institutional Development at Addis Ababa University, on the latest episode of the Techኢት Podcast. 🚀 We discussed his background, Natural Language Processing—especially in local languages—trending AI topics, and institutional matters at Addis Ababa University. Dr. Wondwossen also advised young students in tech. Don’t miss this fascinating discussion about Ethiopia's tech future! 🌍

🎧 Watch now: YouTube link

@Techinethio
#TechPodcast #AI #DigitalTransformation #NLP #EthiopianTech #Innovation #MachineLearning
❤‍🔥6
This media is not supported in your browser
VIEW IN TELEGRAM
🇪🇹 ChatGPT can now speak Ethiopian languages.


Obviously it's not the best but check out when she speaks Amharic, Afan Oromo, and Tigrigna.

The accent is a bit funny but I'm surprised😁
20🤣9👍3🔥2😁1🏆1
If you have worked on vision language models, you have probably faced visual perception as they tend to focus on the wrong areas.

These people in this paper, Attention Prompting on Image for Large Vision-Language Models, basically introduce Attention Prompting on Image (API), where applying a mask to the image enhances LVLMs' visual understanding. This is very cleaver way and simple way.


Full details to find code, paper,demo, huggingface spaces can be found here on their website.

https://yu-rp.github.io/api-prompting/
🔥9