def create_runs():
exp = create_experiment()
with mlflow.start_run() as run:
mlflow.log_metric("m1", 0.1)
run0 = run
with mlflow.start_run() as run:
mlflow.log_metric("m1", 0.2)
return exp, run0, run
def create_nested_runs():
exp = create_experiment()
with mlflow.start_run() as run:
run0 = run
mlflow.log_metric("m1", 10.0)
with mlflow.start_run(nested=True) as run:
run0_min = run
mlflow.log_metric("m1", 1.0)
with mlflow.start_run(nested=True) as run:
run0_max = run
mlflow.log_metric("m1", 100.0)
with mlflow.start_run() as run:
run1 = run
mlflow.log_metric("m1", 20.0)
return exp, run0, run1
def create_simple_run():
exp = create_experiment()
max_depth = 4
model = create_sklearn_model(max_depth)
with mlflow.start_run(run_name="my_run") as run:
mlflow.log_param("max_depth", max_depth)
mlflow.log_metric("rmse", .789)
mlflow.set_tag("my_tag", "my_val")
mlflow.sklearn.log_model(model, "model")
with open("info.txt", "w") as f:
f.write("Hi artifact")
mlflow.log_artifact("info.txt")
mlflow.log_artifact("info.txt", "dir2")
mlflow.log_metric("m1", 0.1)
return exp, run
def create_simple_run():
exp = create_experiment()
mlflow.sklearn.autolog()
X = np.array([[1, 1], [1, 2], [2, 2], [2, 3]])
y = np.dot(X, np.array([1, 2])) + 3
model = LinearRegression()
with mlflow.start_run(run_name="my_run") as run:
mlflow.log_param("p1", "0.1")
mlflow.log_metric("m1", 0.1)
mlflow.set_tag("my_tag", "my_val")
with open("info.txt", "w") as f:
f.write("Hi artifact")
mlflow.log_artifact("info.txt")
mlflow.log_artifact("info.txt", "dir2")
model.fit(X, y)
mlflow.sklearn.log_model(model, "sklearn-model")
return exp, run
def test_no_run_name():
exp = create_experiment()
with mlflow.start_run() as run:
mlflow.log_metric("m1", 0.1)
run2 = client.get_run(run.info.run_id)
assert TAG_RUN_NAME not in run2.data.tags
def create_runs(num_runs):
exp = create_experiment()
for _ in range(0, num_runs):
with mlflow.start_run():
mlflow.log_metric("m1", 0.1)
return exp