This model contains the weights of NExT-GPT covering text-image-video-audio (tiva), which is built upon

• 1. Vicuna-7B with version 0
• 2. ImageBind
• 3. Stable Diffusion with version v1-5.
• 4. AudioLDM with version l-full.
• 5. ZeroScope with version v2_576w.

For more details about the usage of the model, please refer to our code repository.