[NOTE] The following values must be set before you can deploy: HUGGINGFACEHUB_API_TOKEN You can also customize the "MODEL_ID" and "model-volume"
cd GenAIExamples/AudioQnA/kubernetes/intel/cpu/xeon/manifest
export HUGGINGFACEHUB_API_TOKEN="YourOwnToken"
sed -i "s/insert-your-huggingface-token-here/${HUGGINGFACEHUB_API_TOKEN}/g" audioqna.yaml
kubectl apply -f audioqna.yaml
cd GenAIExamples/AudioQnA/kubernetes/intel/hpu/gaudi/manifest
export HUGGINGFACEHUB_API_TOKEN="YourOwnToken"
sed -i "s/insert-your-huggingface-token-here/${HUGGINGFACEHUB_API_TOKEN}/g" audioqna.yaml
kubectl apply -f audioqna.yaml
Make sure all the pods are running, and restart the audioqna-xxxx pod if necessary.
kubectl get pods
curl http://${host_ip}:3008/v1/audioqna -X POST -d '{"audio": "UklGRigAAABXQVZFZm10IBIAAAABAAEARKwAAIhYAQACABAAAABkYXRhAgAAAAEA", "max_tokens":64}' -H 'Content-Type: application/json'