test(hfjobs): add instance type and size to deploy.json

.hfjobs/Dockerfile

@@ -1,9 +1,9 @@
 FROM python:3.12
 
-ENV ENDPOINT_URL="http://host.docker.internal:8000"
 ENV ENDPOINT_API_KEY="undefined"
-ENV ENDPOINT_NUM_SAMPLES=50
-ENV ENDPOINT_TEST_SEED=31415926
+ARG ENDPOINT_URL="http://host.docker.internal:8000"
+ARG ENDPOINT_NUM_SAMPLES=50
+ARG ENDPOINT_TEST_SEED=31415926
 
 RUN --mount=type=bind,source=.hfjobs/requirements.txt,target=/opt/endpoint/tests/requirements.txt \
     python -m pip install -r /opt/endpoint/tests/requirements.txt

.hfjobs/deploy.json

@@ -1,6 +1,7 @@
 {
     "id": "deploy-whisper-vllm-gpu",
     "description": "Build and deploy vLLM based Whisper on Inference Endpoint",
-    "hardware": "l4x1",
+    "instance_type": "nvidia-l4",
+    "instance_size": "x1",
     "model": "openai/whisper-large-v3-turbo"
 }