The pipeline shows the use of super-resolution neural networks to improve the quality and resolution of the input video.
We prepared ONNX versions (using standard torch.onnx.export) of two lightweight NinaSR models ninasr_b0 and ninasr_b1 with different scales (x2/3/4) from the TorchSR repository. The demo uses input video in 360p format and model ninasr_b0 with scale 3 to output video in 1080p format. The model and scale can be changed in the module configuration file.
The demo is prepared for use on dGPU.
Preview:
Tested on platforms:
- Nvidia Turing, Ampere.
git clone https://github.com/insight-platform/Savant.git
cd Savant
git lfs pull
./utils/check-environment-compatibleNote: Ubuntu 22.04 runtime configuration guide helps to configure the runtime to run Savant pipelines.
The demo uses models that are compiled into TensorRT engines the first time the demo is run. This takes time. Optionally, you can prepare the engines before running the demo by using the command:
# you are expected to be in Savant/ directory
./scripts/run_module.py --build-engines samples/super_resolution/module.yml# you are expected to be in Savant/ directory
docker compose -f samples/super_resolution/docker-compose.x86.yml up
# open 'rtsp://127.0.0.1:554/stream/video-super-resolution' in your player
# or visit 'http://127.0.0.1:888/stream/video-super-resolution/' (LL-HLS)
# Ctrl+C to stop running the compose bundle