- Видео 4
- Просмотров 3 464
Together AI
США
Добавлен 28 мар 2024
The official channel of Together AI (together.ai).
Together Talks | Ep 2: Chip Huyen on GPUs & ML Systems Design
Welcome to Together Talks - a series of conversations with researchers and industry leaders on generative AI, brought to you by Together AI.
We’re excited to announce the release of our second episode, featuring a conversation between our Chief Scientist, Tri Dao, and best-selling author and VP of AI & Open Source at Voltron Data, Chip Huyen. Tune in as they dive deep into the world of GPUs and Machine Learning Systems Design!
Learn more at together.ai and follow us on X @togethercompute. Stay tuned for more Together Talks episodes!
____________________________
Chapters
0:00 Introduction: Background & Teaching
7:58 Designing Machine Learning Systems
13:18 Founding Companies [Claypot, Voltron]
18:...
We’re excited to announce the release of our second episode, featuring a conversation between our Chief Scientist, Tri Dao, and best-selling author and VP of AI & Open Source at Voltron Data, Chip Huyen. Tune in as they dive deep into the world of GPUs and Machine Learning Systems Design!
Learn more at together.ai and follow us on X @togethercompute. Stay tuned for more Together Talks episodes!
____________________________
Chapters
0:00 Introduction: Background & Teaching
7:58 Designing Machine Learning Systems
13:18 Founding Companies [Claypot, Voltron]
18:...
Просмотров: 7 396
Видео
Together Talks | Ep 1: Percy Liang on Open Source, AI Evaluations, and the Future of LLMs
Просмотров 1,9 тыс.14 дней назад
Welcome to Together Talks - a series of conversations with researchers and industry leaders on generative AI, brought to you by Together AI. We are thrilled to release our first episode, a conversation between our founders, Percy Liang and Vipul Ved Prakash, on open-source AI, the importance of evaluations, and the future of LLMs. Learn more at together.ai and follow us on X @togethercompute. S...
Together Mixture-Of-Agents explained in 3 minutes
Просмотров 3912 месяца назад
In this video, we'll go over the following: ◆ Explaining how Together Mixture-Of-Agents works ◆ Implementing MoA in just 50 lines of code ◆ Discussing good use cases to leverage MoA ◆ Showing off results from tests we ran
How to run open source AI models on Together AI (Inference + finetuning)
Просмотров 7442 месяца назад
Here's how to run open source LLMs like Llama-3 and finetune them on Together AI.
Thank you
love to hear from you guys
Loved this one!
Thank Tri Dao and Chip Huyen for the great talk. I hope to hear many more talks from both of you in the future.
They are so fun so listen to!
Nghe 2 người bắn tiếng anh mà cứ tưởng mình x2 speed lúc nào k biết 😂😂
Chuẩn bạn, 2 bạn mà bắn chậm rãi rõ ràng hơn tí thì tuyệt vời
2 vietnamese spoke in english =))
They live in the US. Their audience is everyone in the world
Yes their audience is everyone in the world.
chém về tech GPUs, ML ns tiếng Việt k xuông =))
người việt nghe bằng tiếng anh . haha
If you don't know Tri Dao, please educate yourself. He recently gave a good presentation on ml at AI Summit 2024.
I learned a lot from the conversation
Glad to hear!
Great content
Whoa....this holds alot of promise! Good stuff and thanks. First time I hear about this.
Thank you!
Excited for more episodes!
Love this episode!
Can I do this in Nexjs/TypeScript? Please, share a tutorial. Also, is this replacing Vercel AI SDK?
Yes, you can do this in Next.js or TypeScript! You can checkout our docs on TypeScript here: docs.together.ai/docs/introduction Not replacing the Vercel AI SDK, they have different uses.
Beating GPT-4o!! 🔥
Yessir!
Let me know if anyone has any questions!
Do you have a cost calculation comparison between these GPT-4o vs MoA approaches? Would be interesting to know if the extra complexity has financial benefits as well. Thanks for sharing
Nice quick overview.
Thank you!
@@togetherdotai you're welcome. Looking forward to future videos.
Enjoyed recording this, let me know if anyone has any questions!
🙏
Do I need a paid account to fine tune? The page is not showing that right now