Add Resnet-50 CTL FP16 benchmarks

PiperOrigin-RevId: 285568490

Author: Zongwei Zhou

official/benchmark/resnet_ctl_imagenet_benchmark.py

@@ -236,6 +236,17 @@ def benchmark_1_gpu(self):
     FLAGS.batch_size = 128
     self._run_and_report_benchmark()
 
+  def benchmark_1_gpu_fp16(self):
+    """Test Keras model with 1 GPU with tf.keras mixed precision."""
+    self._setup()
+
+    FLAGS.num_gpus = 1
+    FLAGS.distribution_strategy = 'one_device'
+    FLAGS.model_dir = self._get_model_dir('benchmark_1_gpu_fp16')
+    FLAGS.batch_size = 256
+    FLAGS.dtype = 'fp16'
+    self._run_and_report_benchmark()
+
   def benchmark_1_gpu_amp(self):
     """Test Keras model with 1 GPU with automatic mixed precision."""
     self._setup()
@@ -273,6 +284,19 @@ def benchmark_1_gpu_eager(self):
     FLAGS.single_l2_loss_op = True
     self._run_and_report_benchmark()
 
+  def benchmark_1_gpu_fp16_eager(self):
+    """Test Keras model with 1 GPU with fp16 and pure eager mode."""
+    self._setup()
+
+    FLAGS.num_gpus = 1
+    FLAGS.distribution_strategy = 'one_device'
+    FLAGS.model_dir = self._get_model_dir('benchmark_1_gpu_fp16_eager')
+    FLAGS.batch_size = 128
+    FLAGS.dtype = 'fp16'
+    FLAGS.use_tf_function = False
+    FLAGS.single_l2_loss_op = True
+    self._run_and_report_benchmark()
+
   def benchmark_8_gpu(self):
     """Test Keras model with 8 GPUs."""
     self._setup()
@@ -283,6 +307,17 @@ def benchmark_8_gpu(self):
     FLAGS.batch_size = 128 * 8  # 8 GPUs
     self._run_and_report_benchmark()
 
+  def benchmark_8_gpu_fp16(self):
+    """Test Keras model with 8 GPUs with tf.keras mixed precision."""
+    self._setup()
+
+    FLAGS.num_gpus = 8
+    FLAGS.distribution_strategy = 'mirrored'
+    FLAGS.model_dir = self._get_model_dir('benchmark_8_gpu_fp16')
+    FLAGS.batch_size = 256 * 8  # 8 GPUs
+    FLAGS.dtype = 'fp16'
+    self._run_and_report_benchmark()
+
   def benchmark_8_gpu_amp(self):
     """Test Keras model with 8 GPUs with automatic mixed precision."""
     self._setup()