use lstm: True doesn't work with a custom model on a Dict observation space
use lstm: True doesn't work with a custom model on a Dict observation space: a task in MiMo-V2.6-RL-oss: Agentic RL Environments (MiMo RL release). use lstm: True doesn't work with a custom model on a Dict observation space I'm training on a custom env whose observation space is a gym.spaces.Dict…
The task
use_lstm: True doesn't work with a custom model on a Dict observation space I'm training on a custom env whose observation space is a gym.spaces.Dict (a few named sub-fields). I have a custom_model registered via ModelCatalog.register_custom_model(...) that knows how to consume…
Part of XiaomiMiMo/MiMo-V2.6-RL-oss.