policy

Evaluating-the-generalization-ability-of-an-RL-friendly-VLM-CLIP4MC-to-new-tasks-in-Minecraft

RuslanKudinov · PyTorch

or hover any field below to flag it

Overview

Name
Evaluating-the-generalization-ability-of-an-RL-friendly-VLM-CLIP4MC-to-new-tasks-in-Minecraft
Author
RuslanKudinov
Framework
PyTorch
License
unknown
Skill type
manipulation
Evidence level
untested
Task description
The essence of the experiment is to test how well a CLIP4MC model trained on a single set of tasks (e.g., "gathering wood" and "smelting iron") can serve as an internal reward function for an RL agent when performing a new, unfamiliar task (e.g., "building a shelter" and "taming an animal").

Spaces

Action space
other · 0-dim · 0Hz
Observation space
  • type: other

Links

HuggingFace repo
null
Paper (arXiv)
null

Compatible robots

20

Compatible environments

0

No environments list Evaluating-the-generalization-ability-of-an-RL-friendly-VLM-CLIP4MC-to-new-tasks-in-Minecraft yet.

Datasets that reference this policy

0

No datasets reference Evaluating-the-generalization-ability-of-an-RL-friendly-VLM-CLIP4MC-to-new-tasks-in-Minecraft yet.