Deploying Your Custom Model into Riva

This section provides a brief overview on the two main tools used in the deployment process:

1. The build phase using riva-build.

2. The deploy phase using riva-deploy.

Build process

For your custom trained model, refer to the riva-build phase (ASR, NLP, TTS) for your model type. At the end of this phase, you’ll have the Riva Model Intermediate Representation (RMIR) archive for your custom model.

Deploy process

At this point, you already have your RMIR archive. Now, you have two options for deploying this RMIR.

Option 1: Use the Quick Start scripts (riva_init.sh and riva_start.sh) with the appropriate parameters in config.sh.

Option 2: Manually run riva-deploy and then start riva-server with the target model repo.

Option 1: Using Quick Start Scripts to Deploy Your Models (Recommended path)

The Quick Start scripts (riva_init.sh and riva_start.sh) use a particular directory for its operations. This directory is defined by the variable $riva_model_loc specified in config.sh. By default, this is set to use a Docker volume, however, you can specify any local directory to this variable.

By default, the riva_init.sh Quick Start script performs the following:

• Downloads the RMIRs defined and enabled in config.sh from NGC into a subdirectory at $riva_model_loc, specifically $riva_model_loc/rmir.

• Executes riva-deploy for each of the RMIRs at $riva_model_loc/rmir to generate their corresponding Triton Inference Server model repository at $riva_model_loc/models.

When you execute riva_start.sh, it starts with the riva-speech container by mounting this $riva_model_loc directory to /data inside the container.

To deploy your own custom RMIR, or set of RMIRs, you would simply need to place them inside the $riva_model_loc/rmir directory. Ensure that you have defined a directory (that you have access to) in the $riva_model_loc variable in config.sh, since you will need to copy over your RMIRs in its subdirectory. If the subdirectory $riva_model_loc/rmir does not exist, then you’d need to create it and then copy your custom RMIRs there.

If you would like to skip the downloading of the default RMIRs from NGC, then you can set the variable $use_existing_rmirs to true. After your custom RMIRs are inside this $riva_model_loc/rmir directory, you can run riva_init.sh which will execute riva-deploy on your custom RMIRs along with any other RMIRs that are present on that directory and generate the Triton Inferece Server model repo at $riva_model_loc/models.

Next, you can run riva_start.sh and it will start the riva-speech container and load your custom models along with any other models that are present at $riva_model_loc/models. If you only want to load your specific models, ensure that $riva_model_loc/models is empty or the /models directory is not present before you run riva_init.sh. The script riva_init.sh creates the subdirectories /rmir and /models if they are not already there.

For more information about seeing logs and using client containers for testing your models, refer to the Local (Docker) section.

Option 2: Using riva-deploy and the Riva Speech Container (Advanced)

1. Execute riva-deploy. Refer to the Deploy section in Overview for a brief overview on riva-deploy.

   The above command creates the Triton Inference Server model repository at /data/models. If you want to write to any other location other than /data/models, this will require additional manual changes in the embedded artifact directories within the configs within some of the Triton Inference Server model repositories that has model specific artifacts such as class labels. Therefore, stick with /data/models unless you are familiar with Triton Inference Server model repository configurations.

2. Manually start the riva-server Docker container using docker run.

   After the Triton Inference Server model repository for your custom model is generated, start the Riva server on that target repo. The following command assumes you generated the model repo at /data/models.

   docker run -d --gpus 1 --init --shm-size=1G --ulimit memlock=-1 --ulimit stack=67108864 \
           -v /data:/data             \
           -p 50051                            \
           -e '\''CUDA_VISIBLE_DEVICES=0'\''   \
           --name riva-speech                \
           riva-api                          \
           start-riva  --riva-uri=0.0.0.0:50051 --nlp_service=true --asr_service=true --tts_service=true
   

   This command launches the Riva Speech Service API server similar to the Quick Start script riva_start.sh.

   Example output:

   Starting Riva Speech Services
   > Waiting for Triton server to load all models...retrying in 10 seconds
   > Waiting for Triton server to load all models...retrying in 10 seconds
   > Waiting for Triton server to load all models...retrying in 10 seconds
   > Triton server is ready…
   

3. Verify that the servers have started correctly and check that the output of docker logs riva-speech shows:

   I0428 03:14:50.440943 1 riva_server.cc:66] TTS Server connected to Triton Inference Server at 0.0.0.0:8001
   I0428 03:14:50.440943 1 riva_server.cc:66] NLP Server connected to Triton Inference Server at 0.0.0.0:8001
   I0428 03:14:50.440951 1 riva_server.cc:68] ASR Server connected to Triton Inference Server at 0.0.0.0:8001
   I0428 03:14:50.440955 1 riva_server.cc:71] Riva Conversational AI Server listening on 0.0.0.0:50051