HF RL Explorer

I'm using mat64.Dense and getting weird values when I do element-wise ops in place, like d.Sub(d, b) or…

I'm using mat64.Dense and getting weird values when I do element-wise ops in place, like d.Sub(d, b) or…: a task in MiMo-V2.6-RL-oss: Agentic RL Environments (MiMo RL release). Problem Statement I'm using mat64.Dense and getting weird values when I do element-wise ops in place, like d.Sub(d, b) or…

The task

Problem Statement I'm using mat64.Dense and getting weird values when I do element-wise ops in place, like d.Sub(d, b) or d.MulElem(a, d); it feels like the result is getting clobbered when the output is also one of the inputs. I also hit a similar-looking issue with Scale and…

Part of XiaomiMiMo/MiMo-V2.6-RL-oss.