What is Speech Live Portrait Server?
Speech Live Portrait, also known as Speech Animation, animates a person's portrait photo using a driving audio. It also accepts tuning values to control Eye blinks, head movement and gaze of animation. Maxine Speech LivePortrait Server is a Docker image containing end-to-end applications with necessary dependencies that can be easily deployed on public and private clouds and enable client applications to provide the benefits of NVIDIA Maxine Speech LivePortrait algorithms via cloud-based GPU computing.
Prerequisites
NVIDIA Speech Live Portrait microservice supports the Linux x86_64 architecture.
Before you can use Live Portrait microservice, ensure that you meet following prerequisites:
- Ensure you can access and are logged in to NVIDIA NGC. For step-by-step instructions, refer to the NGC Getting Started Guide.
- Verify that you can access to a machine with GPU of any of below architectures:
- sm_70 (e.g. V100)
- sm_75 (e.g. T4)
- sm_86 (e.g. A10, A40)
- sm_89 (e.g. L4, L40)
- Install Docker with support for NVIDIA GPUs.
| Component | Required Software |
|---|---|
| Docker | Docker version later than 19.02 with nvidia-docker installed is required. For non-DGX users, Docker verion 19.03 or later is required. |
| Helm(for kubernetes deployment) | Helm charts 3.x |
| NVIDIA Driver | 535+ |
Limitations
License
By pulling and using Maxine software, you accept the terms and conditions of the Speech Live Portrait License (Under Resources).