Thanks for the feedback. I saw there were also some other people interested in that, and I personally find it a fascinating topic as well. The next article will likely be the upcoming research highlights for October, but then I am planning to follow up with a LoRA one if things go as planned!
On a related note, have you ever thought about doing a podcast series? I get that it’s a different format and a different beast altogether, but thought I’d ask.
That's an interesting suggestion, and to be honest, I have thought about it in the past. However, I currently like all the coding and writing projects I am involved in, and I don't want to give any of that up at the moment to make room/time for a podcast at the moment 😅.
Oct 11, 2023·edited Oct 11, 2023Liked by Sebastian Raschka, PhD
I agree with you here, Sebastian. Focus on, say, 3 key things for 100% instead of trying to do 10 things for 50%. I have seen other AI experts trying to get a podcast from the ground and I find it sub-optimal. But that is personal of course. It is sub-optimal as it is mostly interviews and there are already plenty of those pod casts. In addition, the form factor complete changes: there is a difference between discussing a (complex) topic vs. writing about it. A kind of in-between would be to create a video where you show how it is done, which is essentially your blog post in video format. This is particularly interesting for coding. For instance, when you write about a LORA implementation from scratch, you could also opt to create a walk-through video and show it is done. Just a thought.
Keep up the great work. Your posts are worth every penny!
I have a background in both econometrics and CS, and work in ML in financial industry. Half of my job these days seems to be keeping up with all the developments. Your stuff, Sebastian, is my go to resource. To the point and well presented. You are serving a large community. We appreciate it.
Yes, really good resource. There is a lot of stuff out there, but you get to the important innovations and developments, filter out the chaff, and highlight and explain key pieces and, most importantly, avoid the hype. Well done! Do keep it up!
An Lora implementation from scratch would be interesting
+1 on the lora implementation from scratch .
Perhaps you may also consider a deep-dive on the Lit-gpt implementation as well
Coming here to vote yes on a Lora implementation and explanation.
Thanks for the feedback. I saw there were also some other people interested in that, and I personally find it a fascinating topic as well. The next article will likely be the upcoming research highlights for October, but then I am planning to follow up with a LoRA one if things go as planned!
+2 on the lora implementation from scratch and explanation in a future article to explain how it works in detail.
On a related note, have you ever thought about doing a podcast series? I get that it’s a different format and a different beast altogether, but thought I’d ask.
That's an interesting suggestion, and to be honest, I have thought about it in the past. However, I currently like all the coding and writing projects I am involved in, and I don't want to give any of that up at the moment to make room/time for a podcast at the moment 😅.
I agree with you here, Sebastian. Focus on, say, 3 key things for 100% instead of trying to do 10 things for 50%. I have seen other AI experts trying to get a podcast from the ground and I find it sub-optimal. But that is personal of course. It is sub-optimal as it is mostly interviews and there are already plenty of those pod casts. In addition, the form factor complete changes: there is a difference between discussing a (complex) topic vs. writing about it. A kind of in-between would be to create a video where you show how it is done, which is essentially your blog post in video format. This is particularly interesting for coding. For instance, when you write about a LORA implementation from scratch, you could also opt to create a walk-through video and show it is done. Just a thought.
Keep up the great work. Your posts are worth every penny!
I have a background in both econometrics and CS, and work in ML in financial industry. Half of my job these days seems to be keeping up with all the developments. Your stuff, Sebastian, is my go to resource. To the point and well presented. You are serving a large community. We appreciate it.
Thanks for the kind words!!
+x on the LoRa implementation
btw: great newsletter!!!
Yes, really good resource. There is a lot of stuff out there, but you get to the important innovations and developments, filter out the chaff, and highlight and explain key pieces and, most importantly, avoid the hype. Well done! Do keep it up!
Thank you for curating and compiling a high quality list of AI resources. Your POV on each topic is very helpful!
Thanks, Fabric, glad to hear!
Thank you for the interesting article! +1 on Lora from scratch.
One of the best newsletters to follow, thanks for your great work Sebastian!
Makes sense