@KevinQHLin
Thanks @_akhaliq sharing our work! 🎉We are exciting to introduce VideoCUA 📷A Large-scale video dataset to advance human-level computer-use agent. - Total 55 hours & 6 million frames - 10,000 human-demonstrated tasks - 87 desktop apps, recorded at 30fps 🤗Fully open-source at huggingface https://t.co/u0mmoTw5zD