SQ elements return wrong VM
SQ elements return wrong VM: a task in MiMo-V2.6-RL-oss: Agentic RL Environments (MiMo RL release). SQ elements always have a VM of 1, however: From what I understand, this is also the case for an empty sequence, which is really annoying from a usage point of view. It seems like the value is the…
The task
SQ elements always have a VM of 1, however: From what I understand, this is also the case for an empty sequence, which is really annoying from a usage point of view. It seems like the value is the "sequence", which may or may not contain items. Technically this is…
Part of XiaomiMiMo/MiMo-V2.6-RL-oss.