Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

MLX + OpenCode + Qwen3.5-122B-A10B-4bit on M3 Ultra created a great snake game! Work zero-shot. Video clearly in super fast mode during generation. I generated the prompt using Grok 4.20, it's in the article.

ivanfioravanti's profile picture

Ivan Fioravanti ᯅ

22,131 subscribers

74,659 görüntüleme • 4 ay önce •via X (Twitter)

Bilim & Teknoloji
LIVE

Anya Rossi• Live Now

Private livecam show

0 Yorum

Yorum bulunmuyor

Orijinal gönderinin yorumları burada görünecek

Benzer Videolar

MiniMax M2.1 in 4-bit cruises on an M3 Ultra with mlx-lm. Generated a space invaders game using 5098 tokens at 47.2 tok/sec:
0:23

MiniMax M2.1 in 4-bit cruises on an M3 Ultra with mlx-lm. Generated a space invaders game using 5098 tokens at 47.2 tok/sec:

Awni Hannun

93,826 görüntüleme • 6 ay önce

Step-3.5-Flash in action on MLX with OpenCode on a single (distributed testing in progress!) M3 Ultra to create a snake game! 🔥 6bit quantization. Perfect tool calling. Fast & powerful coding model! Recommended Inference Settings: Temperature: 1.0 Top-p: 0.95 Top-k: 40 🧵
0:27

Step-3.5-Flash in action on MLX with OpenCode on a single (distributed testing in progress!) M3 Ultra to create a snake game! 🔥 6bit quantization. Perfect tool calling. Fast & powerful coding model! Recommended Inference Settings: Temperature: 1.0 Top-p: 0.95 Top-k: 40 🧵

Ivan Fioravanti ᯅ

17,158 görüntüleme • 5 ay önce

Running Minimax M2.1 (MiniMax (official)) with OpenCode (OpenCode) and mlx_lm.server. Works quite well on an M3 Ultra. Once the KV cache is warm the prompt processing is pretty quick. And token generation is very fast.
0:41

Running Minimax M2.1 (MiniMax (official)) with OpenCode (OpenCode) and mlx_lm.server. Works quite well on an M3 Ultra. Once the KV cache is warm the prompt processing is pretty quick. And token generation is very fast.

Awni Hannun

32,329 görüntüleme • 6 ay önce

Qwen3.5 runs quite well in mlx-lm. Awesome that we have a frontier-level hybrid model. The context gets longer but the inference speed and memory use barely change. Here's the Q4 generating a space invaders game on an M3 Ultra. Generated 4,120 tokens at 37.6 tok/s.
0:28

Qwen3.5 runs quite well in mlx-lm. Awesome that we have a frontier-level hybrid model. The context gets longer but the inference speed and memory use barely change. Here's the Q4 generating a space invaders game on an M3 Ultra. Generated 4,120 tokens at 37.6 tok/s.

Awni Hannun

48,673 görüntüleme • 4 ay önce

Qwen 3.5 397B prompt processing on M3 Ultra (with MLX distributed + JACCL) - 3.4× speedup on 4 chips - scaling improves as context increases Really fun to use with opencode; generated a playable Asteroids clone in ~4 minutes (real time, including me playing it a bit).
4:13

Qwen 3.5 397B prompt processing on M3 Ultra (with MLX distributed + JACCL) - 3.4× speedup on 4 chips - scaling improves as context increases Really fun to use with opencode; generated a playable Asteroids clone in ~4 minutes (real time, including me playing it a bit).

Angelos Katharopoulos

23,552 görüntüleme • 4 ay önce

GLM-5 runs with mlx-lm on a single 512GB M3 Ultra in Q4. It's quite good in my initial testing and pretty fast as well. It generated a highly functional space invaders game using 7.1k tokens at 15.4 tok/s and 419GB memory. Thanks to Gökdeniz Gülmez and Tarjei Mandt for the port.
1:11

GLM-5 runs with mlx-lm on a single 512GB M3 Ultra in Q4. It's quite good in my initial testing and pretty fast as well. It generated a highly functional space invaders game using 7.1k tokens at 15.4 tok/s and 419GB memory. Thanks to Gökdeniz Gülmez and Tarjei Mandt for the port.

Awni Hannun

60,599 görüntüleme • 4 ay önce

🔥 DeepSeek R1 671B Q4 - M3 Ultra 512GB with MLX🔥 - 18.43 tokens/sec - Generates a p5js zero-shot, tested at video's end 😱 - Video in real-time, no acceleration! - First test, and I'm blown away! Prompt: "Create an amazing animation using p5js"
3:11

🔥 DeepSeek R1 671B Q4 - M3 Ultra 512GB with MLX🔥 - 18.43 tokens/sec - Generates a p5js zero-shot, tested at video's end 😱 - Video in real-time, no acceleration! - First test, and I'm blown away! Prompt: "Create an amazing animation using p5js"

Ivan Fioravanti ᯅ

46,493 görüntüleme • 1 yıl önce

Kimi K2.5 1T runs on 2 M3 Ultras with mlx-lm in it's native precision. It's actually quite usable. Here it's making a space invaders game. Generated 3856 tokens at 21.9 tok/sec using 350GB per machine. Thanks to Tarjei Mandt for the port.
1:08

Kimi K2.5 1T runs on 2 M3 Ultras with mlx-lm in it's native precision. It's actually quite usable. Here it's making a space invaders game. Generated 3856 tokens at 21.9 tok/sec using 350GB per machine. Thanks to Tarjei Mandt for the port.

Awni Hannun

241,472 görüntüleme • 5 ay önce

A perfect coding model for MLX on Apple silicon.. Qwen delivered again. Runs quite fast on an M3 Ultra. Running the 4-bit quantized with mlx-lm:
0:50

A perfect coding model for MLX on Apple silicon.. Qwen delivered again. Runs quite fast on an M3 Ultra. Running the 4-bit quantized with mlx-lm:

Awni Hannun

186,641 görüntüleme • 11 ay önce

I did it! It works! Using GLM-4.7-4bit with mlx_lm.server and opencode to fix real code locally! 🔥 Here single M3 Ultra 512GB, nex step phase will be 2 using Tensor Parallelism and then apply same changes to exo. Prefill is slow on a single machine, but generation is good.
1:44

I did it! It works! Using GLM-4.7-4bit with mlx_lm.server and opencode to fix real code locally! 🔥 Here single M3 Ultra 512GB, nex step phase will be 2 using Tensor Parallelism and then apply same changes to exo. Prefill is slow on a single machine, but generation is good.

Ivan Fioravanti ᯅ

44,000 görüntüleme • 6 ay önce

A long time coming but new mlx-lm is here with better batching support in the server and Gemma 4. pip install -U mlx-lm Here is a video where a single M3 Ultra serves 5 opencode sessions with Gemma 4 26B that process ~130k tokens in ~1.5 minutes.
2:03

A long time coming but new mlx-lm is here with better batching support in the server and Gemma 4. pip install -U mlx-lm Here is a video where a single M3 Ultra serves 5 opencode sessions with Gemma 4 26B that process ~130k tokens in ~1.5 minutes.

Angelos Katharopoulos

66,128 görüntüleme • 3 ay önce

"122B has to be smarter than 27B" I showed 4 UI components to three Qwen3.5 models and asked them to recreate them from a screenshot alone: - 27B (dense) - 35B-A3B (MoE) - 122B-A10B (MoE) Same screenshot. Same prompt. Same task. Which one do you think nailed it?
0:37

"122B has to be smarter than 27B" I showed 4 UI components to three Qwen3.5 models and asked them to recreate them from a screenshot alone: - 27B (dense) - 35B-A3B (MoE) - 122B-A10B (MoE) Same screenshot. Same prompt. Same task. Which one do you think nailed it?

stevibe

100,929 görüntüleme • 3 ay önce

I spent 5 hours today making OpenClaw even better at Motion Graphics I created the video below in 2 prompts with no asset uploads. It scraped or generated everything from a text prompt. I've added: - better brand scraping - image generation - video generation - music generation and scraping - better transitions on first prompt - basically zero errors - Always outputs a link to the editor you see below so I can open it from any device... (I can create and edit from telegram) Still work to do I will make a video this weekend and release the skill. It's the best remotion skill that exists imo.
0:34

I spent 5 hours today making OpenClaw even better at Motion Graphics I created the video below in 2 prompts with no asset uploads. It scraped or generated everything from a text prompt. I've added: - better brand scraping - image generation - video generation - music generation and scraping - better transitions on first prompt - basically zero errors - Always outputs a link to the editor you see below so I can open it from any device... (I can create and edit from telegram) Still work to do I will make a video this weekend and release the skill. It's the best remotion skill that exists imo.

Riley Brown

120,252 görüntüleme • 4 ay önce

Just for fun, here's what 32 simultaneous long-context generations with Qwen3 Next 80B looks like on an M3 Ultra. Using the new batch generation in mlx-lm. Context size for each is about 5k tokens:
0:41

Just for fun, here's what 32 simultaneous long-context generations with Qwen3 Next 80B looks like on an M3 Ultra. Using the new batch generation in mlx-lm. Context size for each is about 5k tokens:

Awni Hannun

50,272 görüntüleme • 9 ay önce

A 30-second scene created from an image (created by Grok). I animated it for 10 seconds (without a prompt) and then clicked on "Extend Video" twice. Grok decided on the animation, and I think he did a great job. Grok Imagine
0:30

A 30-second scene created from an image (created by Grok). I animated it for 10 seconds (without a prompt) and then clicked on "Extend Video" twice. Grok decided on the animation, and I think he did a great job. Grok Imagine

Déborah

13,083 görüntüleme • 3 ay önce

The new Deep Seek V3 0324 in 4-bit runs at > 20 toks/sec on a 512GB M3 Ultra with mlx-lm!
0:31

The new Deep Seek V3 0324 in 4-bit runs at > 20 toks/sec on a 512GB M3 Ultra with mlx-lm!

Awni Hannun

168,870 görüntüleme • 1 yıl önce

Grok Imagine is a super fast, groundbreaking AI image and video generation experience.
0:20

Grok Imagine is a super fast, groundbreaking AI image and video generation experience.

Grok

917,902 görüntüleme • 2 ay önce

I'm proud to present this video created with Grok. Here are the main steps in the making of this video: 1. I wrote a draft script by hand. 2. I reworked the script with Grok to eliminate inconsistencies and add technical details (what I have them say about the Apollo missions and the stones is accurate). 3. Grok broke it down into scenes. 4. I switched to Grok Imagine's Agent mode: transforming each scene into a prompt. 5. I animated each scene in Agent mode using video, as well as cropping and image selection. 6. I edited it. Here's the result; it's hard to call it AI. I look forward to reading your comments. Grok Imagine Elon Musk
2:42

I'm proud to present this video created with Grok. Here are the main steps in the making of this video: 1. I wrote a draft script by hand. 2. I reworked the script with Grok to eliminate inconsistencies and add technical details (what I have them say about the Apollo missions and the stones is accurate). 3. Grok broke it down into scenes. 4. I switched to Grok Imagine's Agent mode: transforming each scene into a prompt. 5. I animated each scene in Agent mode using video, as well as cropping and image selection. 6. I edited it. Here's the result; it's hard to call it AI. I look forward to reading your comments. Grok Imagine Elon Musk

Déborah

69,706 görüntüleme • 1 ay önce

Grok 4.20 Agents in work
0:53

Grok 4.20 Agents in work

Tech Dev Notes

46,019 görüntüleme • 4 ay önce