Abstract: This work presents a novel approach to zero-shot visual object goal navigation that leverages the ability of visual Large Language Model (vLLM) for finding target in unknown environment. Our ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results