Loading video...

Video Failed to Load

Go Home

DeepSeek-R1 on Mac Studio 192GB 🪄

259,545 views • 1 year ago •via X (Twitter)

12 Comments

Georgi Gerganov's profile picture
Georgi Gerganov1 year ago

The model is quantized by @UnslothAI

Solar Heavy's profile picture
Solar Heavy1 year ago

- Moonlight

Unsloth AI's profile picture
Unsloth AI1 year ago

Looks soo good! 🎇🐋

Daniel Nakov's profile picture
Daniel Nakov1 year ago

come on apple, new studio

Georgi Gerganov's profile picture
Georgi Gerganov1 year ago

M4 studio with 256GB would be the final boss

unfazed.'s profile picture
unfazed.1 year ago

@JudeSkrr il a encore drop

Vasilii Kharitonov 🌍's profile picture
Vasilii Kharitonov 🌍1 year ago

Tried R1 32B 8bit on M4 Pro with 48 GB memory (ggml-org/DeepSeek-R1-Distill-Qwen-32B-Q8_0-GGUF), seems to run great 👍 Memory usage is 43.3/48.0GB, eval time is about 7 tokens per second.

GP's profile picture
GP1 year ago

I need your cli setup 😏

Woody Lu's profile picture
Woody Lu1 year ago

Could you share the noise level (in DB) of the mac when running it locally?

Georgi Gerganov's profile picture
Georgi Gerganov1 year ago

Not sure how to measure it. It only makes noise for very long contexts, otherwise it's completely silent. And when it makes noise, it's not really loud IMO.

Alexander Kvamme's profile picture
Alexander Kvamme1 year ago

What profiling tool are you using? Much cleaner than htop

Georgi Gerganov's profile picture
Georgi Gerganov1 year ago

asitop

Related Videos