Skip to content

GitLab

Commit 64755c7f authored by Zongwei Zhou's avatar Zongwei Zhou Committed by A. Unique TensorFlower
Browse files

Reduce batch size of resnet fp16 eager benchmark to prevent occasional OOM 

PiperOrigin-RevId: 336715941
parent 3e74d7e1
Hide whitespace changes
Inline Side-by-side
Showing with 1 addition and 1 deletion
official/benchmark/resnet_ctl_imagenet_benchmark.py
View file @ 64755c7f
...@@ -296,7 +296,7 @@ class Resnet50CtlBenchmarkBase(CtlBenchmark): ...@@ -296,7 +296,7 @@ class Resnet50CtlBenchmarkBase(CtlBenchmark):
FLAGS.num_gpus = 1 FLAGS.num_gpus = 1
FLAGS.distribution_strategy = 'one_device' FLAGS.distribution_strategy = 'one_device'
FLAGS.model_dir = self._get_model_dir('benchmark_1_gpu_fp16_eager') FLAGS.model_dir = self._get_model_dir('benchmark_1_gpu_fp16_eager')
FLAGS.batch_size = 240 FLAGS.batch_size = 232
FLAGS.dtype = 'fp16' FLAGS.dtype = 'fp16'
FLAGS.use_tf_function = False FLAGS.use_tf_function = False
FLAGS.use_tf_while_loop = False FLAGS.use_tf_while_loop = False
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment