正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Using a trillion parameter LLM and 16 core MacBook Pro to send an email

Beff (e/acc)

247,069 subscribers

208,529 次观看 • 9 个月前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

Grok 5 is currently training at the Supercomputer Colossus 2. Grok 5 trains with 6 trillion parameter and a much larger 10 trillion parameter model.

Grok 5 is currently training at the Supercomputer Colossus 2. Grok 5 trains with 6 trillion parameter and a much larger 10 trillion parameter model.

Testlabor

27,518 次观看 • 2 个月前

Can a VLM see without a vision encoder? We trained one for $100, inspired by Gemma 4 12B. Latency on an M3 Pro MacBook: 112 ms -> 1.1 ms for the image path 30% lower end-to-end image+LLM The architecture is just: patchify the image -> linear projection with pos embeddings -> LLM Writeup:

Can a VLM see without a vision encoder? We trained one for $100, inspired by Gemma 4 12B. Latency on an M3 Pro MacBook: 112 ms -> 1.1 ms for the image path 30% lower end-to-end image+LLM The architecture is just: patchify the image -> linear projection with pos embeddings -> LLM Writeup:

Andi Marafioti

60,016 次观看 • 21 天前

Comparing an old, cheap laptop with a MacBook Pro.... No bias at all!

Comparing an old, cheap laptop with a MacBook Pro.... No bias at all!

Amir

729,395 次观看 • 1 年前

My commission sheet Feel free to shoot a dm or send me an email!

My commission sheet Feel free to shoot a dm or send me an email!

Rebecca Doodles

27,929 次观看 • 1 年前

Tested the new MacBook Pro M4 Pro vs. the Mac mini M1 in LM Studio, running Llama 3.2 3B 4-bit MLX. Results: M1 (8-Core) ------> 25 tok/sec M4 Pro (14-Core) -> 104 tok/sec 🤯 – 4x faster!

Tested the new MacBook Pro M4 Pro vs. the Mac mini M1 in LM Studio, running Llama 3.2 3B 4-bit MLX. Results: M1 (8-Core) ------> 25 tok/sec M4 Pro (14-Core) -> 104 tok/sec 🤯 – 4x faster!

01000010

111,446 次观看 • 1 年前

Send a trillion dollars to Ukraine at this point

Send a trillion dollars to Ukraine at this point

Quality Learing Center

86,828 次观看 • 11 天前

The prompt: "Create a 360 draggable macbook pro"

The prompt: "Create a 360 draggable macbook pro"

John Lindquist

73,672 次观看 • 7 个月前

Razer shows off their new Razer Blade 16 with an RTX 5090 GPU crammed inside its super slim frame! And at 0.59” thick at its thinnest point, it beats even the Macbook Pro! #CES2025 #razer

Razer shows off their new Razer Blade 16 with an RTX 5090 GPU crammed inside its super slim frame! And at 0.59” thick at its thinnest point, it beats even the Macbook Pro! #CES2025 #razer

Newegg

15,138 次观看 • 1 年前

This guy built his own Apple Vision Pro using an iPhone and Meta Quest Pro to recreate EyeSight. Source:

This guy built his own Apple Vision Pro using an iPhone and Meta Quest Pro to recreate EyeSight. Source:

Nathie

41,397 次观看 • 1 年前

I’m givin away a BRAND new M4 MacBook Pro 🔥🔥

Joey Urbinati

51,889 次观看 • 1 年前

Unboxing the M4 Pro MacBook Pro with nano-texture 😮‍💨

Unboxing the M4 Pro MacBook Pro with nano-texture 😮‍💨

Apple Hub

192,391 次观看 • 1 年前

Still need to do some cable management, but this is a huge upgrade to my setup. I put my M3 MacBook Air on a SVALT dock (heatsink + fan) and thermal throttling is completely gone. Now when I’m home it basically performs like a MacBook Pro 🔥

Still need to do some cable management, but this is a huge upgrade to my setup. I put my M3 MacBook Air on a SVALT dock (heatsink + fan) and thermal throttling is completely gone. Now when I’m home it basically performs like a MacBook Pro 🔥

Fabrizio Rinaldi

351,407 次观看 • 1 年前

$IREN CEO Daniel Roberts: "2 hours ago we got an email from a trillion dollar hyperscaler that said: 'we weren't interested in Sweetwater (the 1.4 GW site) and now they are'." 🔥

$IREN CEO Daniel Roberts: "2 hours ago we got an email from a trillion dollar hyperscaler that said: 'we weren't interested in Sweetwater (the 1.4 GW site) and now they are'." 🔥

𝐀𝐠𝐫𝐢𝐩𝐩𝐚 𝐈𝐧𝐯𝐞𝐬𝐭𝐦𝐞𝐧𝐭𝐬

113,710 次观看 • 1 年前

Imagine typing your website's URL, and getting an email template back? All this using your own brand, logo, colors, styling, and built using React + Tailwind. Go try it out:

Zeno Rocha

114,960 次观看 • 1 年前

Holy shit... Microsoft open sourced an inference framework that runs a 100B parameter LLM on a single CPU. It's called BitNet. And it does what was supposed to be impossible. No GPU. No cloud. No $10K hardware setup. Just your laptop running a 100-billion parameter model at human reading speed. Here's how it works: Every other LLM stores weights in 32-bit or 16-bit floats. BitNet uses 1.58 bits. Weights are ternary just -1, 0, or +1. That's it. No floats. No expensive matrix math. Pure integer operations your CPU was already built for. The result: - 100B model runs on a single CPU at 5-7 tokens/second - 2.37x to 6.17x faster than llama.cpp on x86 - 82% lower energy consumption on x86 CPUs - 1.37x to 5.07x speedup on ARM (your MacBook) - Memory drops by 16-32x vs full-precision models The wildest part: Accuracy barely moves. BitNet b1.58 2B4T their flagship model was trained on 4 trillion tokens and benchmarks competitively against full-precision models of the same size. The quantization isn't destroying quality. It's just removing the bloat. What this actually means: - Run AI completely offline. Your data never leaves your machine - Deploy LLMs on phones, IoT devices, edge hardware - No more cloud API bills for inference - AI in regions with no reliable internet The model supports ARM and x86. Works on your MacBook, your Linux box, your Windows machine. 27.4K GitHub stars. 2.2K forks. Built by Microsoft Research. 100% Open Source. MIT License.

Holy shit... Microsoft open sourced an inference framework that runs a 100B parameter LLM on a single CPU. It's called BitNet. And it does what was supposed to be impossible. No GPU. No cloud. No $10K hardware setup. Just your laptop running a 100-billion parameter model at human reading speed. Here's how it works: Every other LLM stores weights in 32-bit or 16-bit floats. BitNet uses 1.58 bits. Weights are ternary just -1, 0, or +1. That's it. No floats. No expensive matrix math. Pure integer operations your CPU was already built for. The result: - 100B model runs on a single CPU at 5-7 tokens/second - 2.37x to 6.17x faster than llama.cpp on x86 - 82% lower energy consumption on x86 CPUs - 1.37x to 5.07x speedup on ARM (your MacBook) - Memory drops by 16-32x vs full-precision models The wildest part: Accuracy barely moves. BitNet b1.58 2B4T their flagship model was trained on 4 trillion tokens and benchmarks competitively against full-precision models of the same size. The quantization isn't destroying quality. It's just removing the bloat. What this actually means: - Run AI completely offline. Your data never leaves your machine - Deploy LLMs on phones, IoT devices, edge hardware - No more cloud API bills for inference - AI in regions with no reliable internet The model supports ARM and x86. Works on your MacBook, your Linux box, your Windows machine. 27.4K GitHub stars. 2.2K forks. Built by Microsoft Research. 100% Open Source. MIT License.

Guri Singh

2,180,357 次观看 • 4 个月前

Send love and luck and create a custom St. Patrick’s Day card to send anywhere in Ireland and around the world, using the online An Post AI Card Creator. We will print and post to Ireland or Worldwide for €4. Try it now for a limited time only: #SendLove

Send love and luck and create a custom St. Patrick’s Day card to send anywhere in Ireland and around the world, using the online An Post AI Card Creator. We will print and post to Ireland or Worldwide for €4. Try it now for a limited time only: #SendLove

An Post

45,261 次观看 • 2 年前

Thrilled to have been part of an incredibly talented team designing and bringing the new iPhone 16 Pro page to life ❤️

Thrilled to have been part of an incredibly talented team designing and bringing the new iPhone 16 Pro page to life ❤️

Anton Repponen

87,527 次观看 • 1 年前