Merge pull request #2 from vbvg2008/master

added pipeline benchmark example

benchmark/pipeline_performance.py

@@ -0,0 +1,15 @@
+from fastestimator.pipeline.static.preprocess import Minmax
+from fastestimator.pipeline.pipeline import Pipeline
+import tensorflow as tf
+import numpy as np
+
+(x_train, y_train), (x_eval, y_eval) = tf.keras.datasets.mnist.load_data()
+x_train = np.expand_dims(x_train, -1)
+
+pipeline = Pipeline(batch_size=32,
+                    feature_name=["x", "y"],
+                    train_data={"x": x_train, "y": y_train},
+                    validation_data= {"x": x_eval, "y": y_eval},
+                    transform_train= [[Minmax()], []])
+
+pipeline.benchmark(num_steps=3000, mode="train")