Hey Richard , very good video . I can do better editing in your videos which can help you to get more engagement in your videos . Pls lmk what do you think ?
@RichardLopezCodes
3 ай бұрын
Hi Divy 👋🏼 and thanks for your comment! I’m open to hearing more about what you have in mind. We can continue talking about this offline. Feel free to reach out to me at: [email protected] 😊
@vernearase3044
3 ай бұрын
I think the main requirement for running the Small Language Model locally on your device is 8 GB memory. It doesn't appear to be the strength of the neural processing unit (NPU) because it will run on the M1 which only has 11 TOPS (Trillion Operations/Second) whereas the A16 Bionic in the iPhone 14 Pro's NPU can run at 17 TOPS - and it _isn't_ supported. The telling difference is that the _all_ the 'M' class SoCs come with at least 8 GB RAM as does the iPhone 15 Pro, where the iPhone 14 Pro only came with 6 GB RAM. All flavors of the iPhone 16 therefore will probably come with at least 8 GB memory, and the A18 Bionic SoC. The A17 Pro and M3 SoCs are all built on TSMC's N3B process node which was custom made for Apple when N3E wasn't ready - and Apple wanted a 3nm process. N3B has lower yields and much more complicated fabrication process, and doesn't fit in with TSMC's future 3nm road map to future nodes like N3P. Apple is _desperate_ to retire N3B because phones or computers built using N3B chips will have to be supported for seven years or so, so M3 is a dead end at this point. (M4 is built on N3E). Once Apple stocks up on enough A17 Pros and M3, M3 Pro, and M3 Max chips for replacement repairs TSMC can retire the N3B process nodes. If Apple runs out of M3 or A17 Pro boards after N3B goes down, they'll probably offer more recent models as replacements because N3B is costing Apple a _lot_ of money.
@RichardLopezCodes
3 ай бұрын
@@vernearase3044 thanks for taking the time to write this detailed comment and add more context! A look at some articles suggests that Apple Intelligence requires roughly 2GB of spare memory to run (based on its 3 Billion parameter model). So I agree that the lack of memory in some iPhone models might be causing limitations. Unfortunately, I will be in the camp that needs an upgrade 😔
@ProductTalksWithMichael
3 ай бұрын
Great breakdown!!
@RichardLopezCodes
3 ай бұрын
Thanks for giving it a watch !
@kiarashkarami444
3 ай бұрын
prefect video but i think your main Chanel problem is thumbnail
@RichardLopezCodes
3 ай бұрын
Appreciate the feedback! I agree, I don't think it's very clickable. I just made a new one and will be running tests to compare it to the original :)
@pedermyhrer3490
4 ай бұрын
Thank you so much for the video! Very informative, and i got right to making my app.
@RichardLopezCodes
4 ай бұрын
Thank you for the nice comment! I'm glad it could help :) I'll be releasing more videos soon, hopefully 🤞🏽
@midi-24
5 ай бұрын
Hi, you should add the python code in your description so others can use it too. Your cheat sheet doc is great. Thanks!
@RichardLopezCodes
5 ай бұрын
Hi! Thanks for letting me know! I just updated the description, but here is the link to the python code too: github.com/rl3020/ChatGptTutorial . Glad the cheat sheet doc is useful :)
Пікірлер