Creates nightly benchmark

This creates a new nightly benchmark action. Right now, it only contains a
single mistral test for a POC. Later, this config file can be replaced with a
more intentional suite to track.

.github/workflows/benchmark-nightly.yml

@@ -0,0 +1,19 @@
+name: Benchmark Nightly
+
+on:
+  pull_request:
+  schedule:
+    - cron: '0 1 * * *'
+
+permissions:
+  id-token: write
+  contents: read
+
+jobs:
+  g5-2xl:
+    uses: ./.github/workflows/instant_benchmark.yml
+    secrets: inherit
+    with:
+      running_template: tests/benchmark/nightly/g5-2xl.txt
+      instance: g5.2xlarge
+      record: cloudwatch

.github/workflows/instant_benchmark.yml

@@ -34,6 +34,22 @@ on:
           - none
           - table
           - cloudwatch
+  workflow_call:
+    inputs:
+      running_template:
+        required: true
+        type: string
+      instance:
+        required: true
+        type: string
+      container:
+        required: false
+        type: string
+        default: ''
+      record:
+        required: false
+        type: string
+        default: 'none'
 
 permissions:
   id-token: write

tests/benchmark/nightly/g5-2xl.txt

@@ -0,0 +1,16 @@
+[test_name]
+mistral
+[container]
+deepjavalibrary/djl-serving:0.27.0-deepspeed
+[vars]
+CONCURRENCY={1,16,32}
+[serving_properties]
+engine=Python
+option.rolling_batch=vllm
+option.model_id=mistralai/Mistral-7B-v0.1
+[aws_curl]
+TOKENIZER=mistralai/Mistral-7B-v0.1 ./awscurl -c $CONCURRENCY -N 10 \
+-X POST http://127.0.0.1:8080/invocations   \
+--connect-timeout 60   -H "Content-type: application/json"   \
+-d '{"inputs":"The new movie that got Oscar this year","parameters":{"max_new_tokens":256, "do_sample":true}}'   \
+-t -o /tmp/output.txt