Відео

How can I Change From Cloudflare Free Account Nameservers to the Nameserver Of My Website Host Server, And Move DNS Control From Cloudflare To My Hosting CPanel?

70b one is 4 bit quantized?

I just ran an 8B Hermes3 model on a 9 year old 2016 MacBook Pro (Core i7 chip with just 2 cores, and 16gb of RAM). It was quite slow though, about 2 words per second, but it worked!

You need an external NVIDIA Project Digits if you want high performance for big LLMs. It will be available for sale in Mai, 2025.

thanks bro you explain in simple words

I wonder how much performance of running llms could improve if opting to m4 pro with 20 GPU..

The problem with these types of video is how dependent they are on your own audio setup. It's really something that either needs hard decibel numbers or to be seen in person.

Follow up. How is that noise level treating you

Why did you go with the 12/16 instead of the 14/20 spec? I’m debating which one to get as a dev who needs to run Parallels with Windows for some Windows desktop development.

after watching your clip, i wonder how the performance of m4 pro 14cpu 20gpu would turn out.

Have you tried something like Private LLM for comparison in terms of speed?

Have you tried MLX versions of llama to see if you're able to go bigger using Apple GPU optimized models?

As a 16" mbp 2019 owner, I think that would be nice to run the same test with low power mode on, just to evaluate the best tradeoff between performance and fan noise, thanks for the video!

Knowing now how the llms perform locally, would you have gotten a max and at least a 64gb MBP instead?

Fan is loud comes on straight away

Between M3 Max (30 cores GPU and only 36GB RAM) and M4 Pro with 48GB RAM, which would be faster?

Do the bigger models require more ram only or ram + gpu?

Can you run any of these on the machine’s Neural Engine cores?

where 405b?

First off I want to say you do an excellent job with your videos. As someone 100% brand new to LLM’s (I sold my 2019 MacBook so don’t even have a computer right now and am deciding which model) I’m wondering if you have any beginner courses of what software to install, etc. to get up and running. I mean from absolute beginner to the point you are at when running these models? Any help or suggestions would be greatly appreciated. Happy Holidays and best wishes for a fabulous 2025! 🎉 🙏

I have a MBP M1 with 16gb RAM and running 70B Llama 3.1 and it's fast

As if you could do much with 70B llama. Try duck duck go chat. It’s mostly useless for complex tasks

Great video, does anyone know what widget or application he is using to monitor the temperature and fan speed of the mac?

What about Mac mini m4 pro with 64 GB?

Did you maximize VRAM allocation to get the whole model into memory? If not, that will explain why it's so slow.

!4GB-8GB Goes to deal with screen and GPU

Thanks for the review. Can you confirm that the MBP is generally quite as would be a fanless MacBook Air?

You should you LM Studio, it will be faster

Excuse me, brother, I want to communicate with you. It is not an important matter. I have a problem. I hope you can help me. I have a problem with Cloudflare.

Excuse me, brother, I want to communicate with you. It is not an important matter. I have a problem. I hope you can help me. I have a problem with Cloudflare.

Im thinking to buy 2019 for like 350-400$. Does it worth it?

Llama 3.3 70b... It runs fine on my 40 core MBP M4 Max with 48GB.

That was fun. I found it was easier to print to pdf then upload to GPT Here is the prompt: Analyze the given website engagement metrics, focusing on the types of sites visited, their frequency, and overall patterns of usage. Based on this data, infer the likely profile of the user, including their profession, interests, and potentially their age range or life stage. Consider the high engagement with professional, technical, and AI-related sites like Google Docs, LinkedIn, Hugging Face, and WordLift, alongside personal productivity tools like Google Calendar and Duolingo. Suggest what kind of work the user does, their level of expertise in technology or a specific domain, and their main areas of focus. Be concise and provide a well-rounded user persona. Have fun!

0:52 Are you sure it's 90W? I thought the binned 12-core accepts 70W? That's actually one of the reasons why I want to buy the 12-core instead of the 14-core - to make sure the TDP is lower. Thanks for the review.

Which llm is best when you compare those llm and which one gives 100% accurate results?

Wasn’t there a mode to throttle performance if you don’t want to get it too hot ?

Hmm I have 64GB on a 7 year old GPU and it runs fine. But 48GB on a mac is basically 96 on a PC so it shouldn't matter

its giving windows vibe ehheheeheh

so should i buy imac m4, mac mini m4, or macbook pro m4 to play some roblox, minecraft and do some video editing? - tim:)

The model has to be entirely in real memory.

can't we just autoscan the DNS records?

Got one, I don't need earphones when making full load (comparing to windows when you need them when open browser)

my registrar asks also the ip address for every name server, so i cannot proceed :/ may u help me in some way? thank you in advance

11:30 Model Results - need a MBP Max with 64GB ram minimum to run llama3.1 and bigger models

this is a cool video. i'm looking to buy a macbook for software development bit i'm unsure what spec to get. i want the fans to turn on as little as possible as i live in in dusty environment. would you recommend the M4 24GB, M4 32GB, M4 Pro 12c 24GB or M4 Pro 12c 48 ram

Try LM Studio.

I try to find this application in the App Store, but I can't find it. May it be due to the fact that my computer is a Macbook Air M1, and maybe they don't give access to this application for computers that don't have a fan? (Can you answer me, please?)

The llama 3.1 is still faster than writing the 500 words story ourselves 🙂

I can run llama3.2-vision using Ollama on my 16GB M1 Macbook Air without any issues. Having the same feelings about the Phi, Mistral7B and the llama3.2. Hope the new Ministral 8B is making it to Ollama. ps here my llama3.2-vision stats using your prompt for comparision: total duration: 1m22.165174917s load duration: 29.836ms prompt eval count: 133 token(s) prompt eval duration: 2.959s prompt eval rate: 44.95 tokens/s eval count: 715 token(s) eval duration: 1m19.173s eval rate: 9.03 tokens/s

I just seen you post this! thanks so much, I havnt watched it yet, but people asked for it and you took the time to do it. Thank you! -----Watching now.