Dario: “With open source you can see the source code. Here we can’t see inside the model.” My 0.646 t/s MiMo-2.6-Pro crammed thru 128GB of Jetson AGX Thor: “Yeah, well, that’s just, like, your opinion, man.”
英文标题：Dario: “With open source you can see the source code. Here we can’t see inside the model.” My 0.646 t/s MiMo-2.6-Pro crammed thru 128GB of Jetson AGX Thor: “Yeah, well, that’s just, like, your opinion, man.”
Tracked my first run of MiMo Pro on a Jetson AGX Thor at a scorching 0.646 tokens/sec,.. but it DID technically load and run, so it counts. Every decoder layer, routed expert, SSD read (..a good bit more) traced and replayable step by step. Here we can't what-now?.. it's ~566GB of model (FP8
中文
Tracked my first run of MiMo Pro on a Jetson AGX Thor at a scorching 0.646 tokens/sec,.. but it DID technically load and run, so it counts. Every decoder layer, routed expert, SSD read (..a good bit more) traced and replayable step by step. Here we can't what-now?.. it's ~566GB of model (FP8
中文全文翻译正在补齐，可先查看下方英文原文。
English / 原文
Tracked my first run of MiMo Pro on a Jetson AGX Thor at a scorching 0.646 tokens/sec,.. but it DID technically load and run, so it counts. Every decoder layer, routed expert, SSD read (..a good bit more) traced and replayable step by step. Here we can't what-now?.. it's ~566GB of model (FP8