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。

First kinda working implementation of GLM 5.2 in DwarfStar. Will take some time to be good enough, but it is a promising start. 433 GB GGUF file. M3 Ultra 512GB.

antirez

75,380 subscribers

69,508 次观看 • 1 个月前 •via X (Twitter)

游戏教育科学技术

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

I didn't expect DeepSeek v4 PRO (not Flash) to run well on the Mac Studio M3 Ultra with 512GB of RAM. This is 2 bit quantized with the same DwarfStar recipe used for Flash. 433GB GGUF file. 130 t/s prefill, 13 t/s generation. Prefill in the video is low because small prompt.

I didn't expect DeepSeek v4 PRO (not Flash) to run well on the Mac Studio M3 Ultra with 512GB of RAM. This is 2 bit quantized with the same DwarfStar recipe used for Flash. 433GB GGUF file. 130 t/s prefill, 13 t/s generation. Prefill in the video is low because small prompt.

antirez

171,478 次观看 • 2 个月前

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.

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 次观看 • 7 个月前

GLM-5.2 8bit running on two M3 Ultra 512GB with MLX distributed? Here it is! 🚀 Decode speed: 17.9 tokens/sec 🔥 Memory used: ~ 760GB 👀 Again keep in mind it's a preliminary PR by super Pedro Cuenca still a WIP!

GLM-5.2 8bit running on two M3 Ultra 512GB with MLX distributed? Here it is! 🚀 Decode speed: 17.9 tokens/sec 🔥 Memory used: ~ 760GB 👀 Again keep in mind it's a preliminary PR by super Pedro Cuenca still a WIP!

Ivan Fioravanti ᯅ

34,175 次观看 • 1 个月前

MLX GLM 5.2 Distributed on two M3 Ultra 512GB 🔥 One M3 Ultra: 18.8 tokens/sec Two M3 Ultra: 23.4 tokens/sec Context: - PR by Pedro Cuenca is still open and probably there is room for improvement: - basic generation test to measure decoding performance here, I will do a full context benchmarking once PR is more mature - nvfp4 quantization used - Video alternates standard speed and x20, with one Mac first and distributed later. Enjoy! 🙌🏻

MLX GLM 5.2 Distributed on two M3 Ultra 512GB 🔥 One M3 Ultra: 18.8 tokens/sec Two M3 Ultra: 23.4 tokens/sec Context: - PR by Pedro Cuenca is still open and probably there is room for improvement: - basic generation test to measure decoding performance here, I will do a full context benchmarking once PR is more mature - nvfp4 quantization used - Video alternates standard speed and x20, with one Mac first and distributed later. Enjoy! 🙌🏻

Ivan Fioravanti ᯅ

87,805 次观看 • 1 个月前

Took the good work of the communtiy of DwarfStar and consolidating the Strix Halo support. It looks very good. More QA in the next days and the final merge soon.

Took the good work of the communtiy of DwarfStar and consolidating the Strix Halo support. It looks very good. More QA in the next days and the final merge soon.

antirez

17,303 次观看 • 2 个月前

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.

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 次观看 • 6 个月前

GLM 5.2, Q2_K routed experts (effectively ~2.6 bits) running with SSD streaming on an M5 Max 128GB computer.

GLM 5.2, Q2_K routed experts (effectively ~2.6 bits) running with SSD streaming on an M5 Max 128GB computer.

antirez

318,383 次观看 • 1 个月前

MPJ: "Some of the shots and some of the ways they (the Nuggets) probably could have used me we had enough good players to where we could create a good shot in a different way...it is good to be able to kinda show what I'm capable of"

MPJ: "Some of the shots and some of the ways they (the Nuggets) probably could have used me we had enough good players to where we could create a good shot in a different way...it is good to be able to kinda show what I'm capable of"

Oh No He Didn't

158,302 次观看 • 7 个月前

DS4 running on DGX Spark (GB10 / CUDA), private branch for now. 12 tokens/sec, the memory bandwidth is limited in this system, at 270GB/sec. But prefill is ways more alighed to M3 Max at ~200 t/s. I'll release when more mature, but it is almost sure that it will get merged.

DS4 running on DGX Spark (GB10 / CUDA), private branch for now. 12 tokens/sec, the memory bandwidth is limited in this system, at 270GB/sec. But prefill is ways more alighed to M3 Max at ~200 t/s. I'll release when more mature, but it is almost sure that it will get merged.

antirez

84,127 次观看 • 3 个月前

Conservative prosecutors are resigning instead of carrying out Trump DOJ political orders. “You will eventually find someone who is enough of a fool, or enough of a coward, to file your motion. But it was never going to be me.” Republicans in Congress—STAND UP.

Conservative prosecutors are resigning instead of carrying out Trump DOJ political orders. “You will eventually find someone who is enough of a fool, or enough of a coward, to file your motion. But it was never going to be me.” Republicans in Congress—STAND UP.

Amy Klobuchar

385,534 次观看 • 1 年前

Zavier Gozo is the best attacking talent I’ve seen out of MLS in quite some time. Believe the hype; he’s good enough to start in a Champions League team.

Zavier Gozo is the best attacking talent I’ve seen out of MLS in quite some time. Believe the hype; he’s good enough to start in a Champions League team.

Zach Lowy

86,496 次观看 • 3 个月前

Ran two prompts in Kimi K2.7 Code HighSpeed, Kimi K2.7 Code, and GLM 5.2 HighSpeed is 2.5 to 4.2 times faster than Kimi K2.7 Code while producing comparable quality 4.5 to 15 times faster than GLM 5.2, and half the price (or less) These are simple tests, so the pricing difference is small and speed difference could diverge a lot But definitely a hell of a lot faster with good output

Ran two prompts in Kimi K2.7 Code HighSpeed, Kimi K2.7 Code, and GLM 5.2 HighSpeed is 2.5 to 4.2 times faster than Kimi K2.7 Code while producing comparable quality 4.5 to 15 times faster than GLM 5.2, and half the price (or less) These are simple tests, so the pricing difference is small and speed difference could diverge a lot But definitely a hell of a lot faster with good output

GMI Cloud

38,174 次观看 • 1 个月前

“I expect you will eventually find someone who is enough of a fool, or enough of a coward, to file your motion. But it was never going to be me.” A career prosecutor stood up to Emil Bove’s lawlessness. Republicans in the Senate should do the same and reject his nomination.

“I expect you will eventually find someone who is enough of a fool, or enough of a coward, to file your motion. But it was never going to be me.” A career prosecutor stood up to Emil Bove’s lawlessness. Republicans in the Senate should do the same and reject his nomination.

Amy Klobuchar

261,408 次观看 • 1 年前

Bryce Lance - WR - NDSU If GB is going to take a WR. He is going to be very packery. (Would guess he will be a tier 1 fit) He never got to play with Christian Watson at NDSU. But maybe he can in GB.

Bryce Lance - WR - NDSU If GB is going to take a WR. He is going to be very packery. (Would guess he will be a tier 1 fit) He never got to play with Christian Watson at NDSU. But maybe he can in GB.

Jacob Morley

100,354 次观看 • 6 个月前

I was a bit suspicious of the claim, but GLM-5.2 is pretty good at designing stuff. Obviously not at the level of a professional designer, but it has that Opus-level quality. Great at: - games - landing pages - HTML artifacts - 3D worlds Wish I had Fable 5 to compare with.

I was a bit suspicious of the claim, but GLM-5.2 is pretty good at designing stuff. Obviously not at the level of a professional designer, but it has that Opus-level quality. Great at: - games - landing pages - HTML artifacts - 3D worlds Wish I had Fable 5 to compare with.

elvis

18,976 次观看 • 1 个月前

Much better at 17 tokens/sec in M3 Max, fusing a few operations. 128gb m3 max.

Much better at 17 tokens/sec in M3 Max, fusing a few operations. 128gb m3 max.

antirez

25,153 次观看 • 3 个月前

$GLM 5.2 is the best Open Model in the world, and many are saying it's almost as good as Opus 4.8 at a fraction of the cost. Here's exactly how to set it up in Cursor using OpenRouter. Complete instructions below 👇$

GLM 5.2 is the best Open Model in the world, and many are saying it's almost as good as Opus 4.8 at a fraction of the cost. Here's exactly how to set it up in Cursor using OpenRouter. Complete instructions below 👇

Agent Native

21,432 次观看 • 1 个月前

This was the start of Messi hatred for a lot of Nigerians, but some of us were objective enough to see how Good Messi truly is and moved on

This was the start of Messi hatred for a lot of Nigerians, but some of us were objective enough to see how Good Messi truly is and moved on

Leonard Onochie

132,527 次观看 • 2 个月前

Discord is working on voice filters that will allow you to change your voice in VCs. Some of these filters will only be available to Discord Nitro subscribers, but it is not yet clear which ones.

Discord is working on voice filters that will allow you to change your voice in VCs. Some of these filters will only be available to Discord Nitro subscribers, but it is not yet clear which ones.

Discord Previews

86,665 次观看 • 1 年前