Lipsync real3dportrait worker

[pschroedl]

This pull request introduces an implementation of the Text to Speech pipeline within the AI worker.

This implements the real3dportrait repo as a pipeline, in addition to the basic functionality of the Parler Text-to-Speech model (https://huggingface.co/parler-tts/parler-tts-large-v1)

This PR follows the pattern of adding pipelines dependencies with minimal dependency changes to the ai-runner requirements.txt by using an additional dockerfile definition. Example usage for local development and testing:

cd ai-worker/runner
docker build -t livepeer/ai-runner:base .
docker build -f docker/Dockerfile.lipsync -t livepeer/ai-runner:lipsync .

docker run --name lipsync-worker -e PIPELINE=lipsync-e MODEL_ID=parler-tts/parler-tts-large-v1 -e HUGGINGFACE_TOKEN={token} --gpus all -p 8000:8000 -v ~/.lpData/models:/models livepeer/ai-runner:lipsync