chore: import upstream snapshot with attribution

release/ml_user_tests/train/compute_tpl_aws.yaml

@@ -0,0 +1,15 @@
+cloud_id: {{env["ANYSCALE_CLOUD_ID"]}}
+region: us-west-2
+
+max_workers: 2
+
+head_node_type:
+    name: head_node
+    instance_type: g6.12xlarge
+
+worker_node_types:
+    -   name: worker_node
+        instance_type: g6.12xlarge
+        min_workers: 2
+        max_workers: 2
+        use_spot: false

release/ml_user_tests/train/compute_tpl_gce.yaml

@@ -0,0 +1,17 @@
+cloud_id: {{env["ANYSCALE_CLOUD_ID"]}}
+region: us-west1
+allowed_azs:
+    - us-west1-b
+
+max_workers: 2
+
+head_node_type:
+    name: head_node
+    instance_type: n1-standard-32-nvidia-tesla-t4-2
+
+worker_node_types:
+    -   name: worker_node
+        instance_type: n1-standard-32-nvidia-tesla-t4-2
+        min_workers: 2
+        max_workers: 2
+        use_spot: false

release/ml_user_tests/train/driver_requirements.txt

@@ -0,0 +1,8 @@
+# Make sure the driver versions are the same as cluster versions.
+# The cluster uses ray-ml Docker image.
+# ray-ml Docker image installs dependencies from ray/python/requirements/ml/ directory.
+# We constrain on these requirements file so that the same versions are installed.
+-c ../../../python/requirements/ml/dl-cpu-requirements.txt
+
+torch
+tensorflow

release/ml_user_tests/train/driver_setup.sh

@@ -0,0 +1,5 @@
+#!/bin/bash
+
+cd "${0%/*}" || exit 1
+
+pip install -U -r ./driver_requirements.txt

release/ml_user_tests/train/train_tensorflow_mnist_test.py

@@ -0,0 +1,34 @@
+import json
+import os
+import time
+
+import ray
+from ray.train.examples.tf.tensorflow_mnist_example import train_tensorflow_mnist
+
+if __name__ == "__main__":
+    start = time.time()
+
+    addr = os.environ.get("RAY_ADDRESS")
+    job_name = os.environ.get("RAY_JOB_NAME", "train_tensorflow_mnist_test")
+
+    if addr is not None and addr.startswith("anyscale://"):
+        ray.init(address=addr, job_name=job_name)
+    else:
+        ray.init(address="auto")
+
+    train_tensorflow_mnist(
+        num_workers=6, use_gpu=True, epochs=20, storage_path="/mnt/cluster_storage"
+    )
+
+    taken = time.time() - start
+    result = {
+        "time_taken": taken,
+    }
+    test_output_json = os.environ.get(
+        "TEST_OUTPUT_JSON", "/tmp/train_tensorflow_mnist_test.json"
+    )
+
+    with open(test_output_json, "wt") as f:
+        json.dump(result, f)
+
+    print("Test Successful!")

release/ml_user_tests/train/train_torch_linear_test.py

@@ -0,0 +1,33 @@
+import json
+import os
+import time
+
+import ray
+
+from ray.train.examples.pytorch.torch_linear_example import train_linear
+
+if __name__ == "__main__":
+    start = time.time()
+
+    addr = os.environ.get("RAY_ADDRESS")
+    job_name = os.environ.get("RAY_JOB_NAME", "train_torch_linear_test")
+
+    if addr is not None and addr.startswith("anyscale://"):
+        ray.init(address=addr, job_name=job_name)
+    else:
+        ray.init(address="auto")
+
+    results = train_linear(
+        num_workers=6, use_gpu=True, epochs=20, storage_path="/mnt/cluster_storage"
+    )
+
+    taken = time.time() - start
+    result = {"time_taken": taken}
+    test_output_json = os.environ.get(
+        "TEST_OUTPUT_JSON", "/tmp/train_torch_linear_test.json"
+    )
+
+    with open(test_output_json, "wt") as f:
+        json.dump(result, f)
+
+    print("Test Successful!")