Search for a command to run...
Combining inherent knowledge of vision-language models with unsupervised domain adaptation through strong-weak guidance