resnet50 training rasie "tensorflow.python.framework.errors_impl.InvalidArgumentError: Retval[0] does not have value" on aarch64

[Danliran]

System information

Have I written custom code (as opposed to using a stock example script provided in TensorFlow):no OS Platform and Distribution (e.g., Linux Ubuntu 16.04):ubuntu 18.04 Mobile device (e.g. iPhone 8, Pixel 2, Samsung Galaxy) if the issue happens on mobile device: no TensorFlow installed from (source or binary): source TensorFlow version (use command below):1.15.0 Python version:2.7 Bazel version (if compiling from source):0.26.0 GCC/Compiler version (if compiling from source):7.4.0 CUDA/cuDNN version: no GPU model and memory:no Describe the current behavior

I used tensorflow/benchmark to train resnet50 by tf_cnn_benchmarks.py on my aarch64 platform. command line: python tf_cnn_benchmarks.py --device=cpu --data_format=NHWC --optimizer=sgd --distortions=false --variable_update=replicated --data_dir=/opt/imagenet/tf_train --data_name=imagenet --model=resnet50 --batch_size=32 --train_dir=/opt/data/resnet50_ckpt/cpu_v1/ --num_epochs=10--save_model_steps=20

Describe the expected behavior After training thousands batchs raise "tensorflow.python.framework.errors_impl.InvalidArgumentError: Retval[0] does not have value"

INFO:tensorflow:Error reported to Coordinator: <class 'tensorflow.python.framework.errors_impl.InvalidArgumentError'>, Retval[0] does not have value I0605 18:58:49.349148 281473050139600 coordinator.py:224] Error reported to Coordinator: <class 'tensorflow.python.framework.errors_impl.InvalidArgumentError'>, Retval[0] does not have value Traceback (most recent call last): File "tf_cnn_benchmarks.py", line 73, in app.run(main) # Raises error on invalid flags, unlike tf.app.run() File "/root/.env/tf2.0.0/lib/python2.7/site-packages/absl/app.py", line 299, in run _run_main(main, args) File "/root/.env/tf2.0.0/lib/python2.7/site-packages/absl/app.py", line 250, in _run_main sys.exit(main(argv)) File "tf_cnn_benchmarks.py", line 68, in main bench.run() File "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", line 1880, in run return self._benchmark_train() File "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", line 2085, in _benchmark_train return self._benchmark_graph(result_to_benchmark, eval_build_results) File "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", line 2294, in _benchmark_graph is_chief, summary_writer, profiler) File "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", line 2430, in benchmark_with_session collective_graph_key=collective_graph_key) File "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", line 869, in benchmark_one_step results = sess.run(fetches, options=run_options, run_metadata=run_metadata) File "/root/.env/tf2.0.0/lib/python2.7/site-packages/tensorflow_core/python/client/session.py", line 956, in run run_metadata_ptr) File "/root/.env/tf2.0.0/lib/python2.7/site-packages/tensorflow_core/python/client/session.py", line 1180, in _run feed_dict_tensor, options, run_metadata) File "/root/.env/tf2.0.0/lib/python2.7/site-packages/tensorflow_core/python/client/session.py", line 1359, in _do_run run_metadata) File "/root/.env/tf2.0.0/lib/python2.7/site-packages/tensorflow_core/python/client/session.py", line 1384, in _do_call raise type(e)(node_def, op, message) tensorflow.python.framework.errors_impl.InvalidArgumentError: Retval[0] does not have value

strace log about training:

openat(AT_FDCWD, "tf_cnn_benchmarks.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=2476, ...}) = 0 read(3, "# Copyright 2017 The TensorFlow "..., 4096) = 2476 write(2, " ", 4 ) = 4 write(2, "app.run(main) # Raises error on"..., 68app.run(main) # Raises error on invalid flags, unlike tf.app.run() ) = 68 close(3) = 0 write(2, " File "/root/.env/tf1.15/lib/py"..., 85 File "/root/.env/tf1.15/lib/python2.7/site-packages/absl/app.py", line 299, in run ) = 85 openat(AT_FDCWD, "/root/.env/tf1.15/lib/python2.7/site-packages/absl/app.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=14864, ...}) = 0 read(3, "# Copyright 2017 The Abseil Auth"..., 4096) = 4096 read(3, "ilar to HelpfullFlag, but genera"..., 4096) = 4096 read(3, "\n logging.error(traceback"..., 4096) = 4096 write(2, " ", 4 ) = 4 write(2, "_run_main(main, args)\n", 22_run_main(main, args) ) = 22 close(3) = 0 write(2, " File "/root/.env/tf1.15/lib/py"..., 91 File "/root/.env/tf1.15/lib/python2.7/site-packages/absl/app.py", line 250, in run_main ) = 91 openat(AT_FDCWD, "/root/.env/tf1.15/lib/python2.7/site-packages/absl/app.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=14864, ...}) = 0 read(3, "# Copyright 2017 The Abseil Auth"..., 4096) = 4096 read(3, "ilar to HelpfullFlag, but genera"..., 4096) = 4096 write(2, " ", 4 ) = 4 write(2, "sys.exit(main(argv))\n", 21sys.exit(main(argv)) ) = 21 close(3) = 0 write(2, " File "tf_cnn_benchmarks.py", l"..., 48 File "tf_cnn_benchmarks.py", line 68, in main ) = 48 openat(AT_FDCWD, "tf_cnn_benchmarks.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=2476, ...}) = 0 read(3, "# Copyright 2017 The TensorFlow "..., 4096) = 2476 write(2, " ", 4 ) = 4 write(2, "bench.run()\n", 12bench.run() ) = 12 close(3) = 0 write(2, " File "/home/tf/benchma"..., 99 File "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", line 1880, in run ) = 99 openat(AT_FDCWD, "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=164060, ...}) = 0 read(3, "# Copyright 2017 The TensorFlow "..., 4096) = 4096 read(3, "\n# the forward-only option, wh"..., 4096) = 4096 read(3, " 'The autotune thresh"..., 4096) = 4096 read(3, " 'the number of s"..., 4096) = 4096 read(3, "ed, after the graph has been par"..., 4096) = 4096 read(3, " maximum number of GPU Ops that "..., 4096) = 4096 read(3, "or the MultiDeviceIterator that "..., 4096) = 4096 read(3, "INE_integer('kmp_settings', 1,\n "..., 4096) = 4096 read(3, " 'collective_all_redu"..., 4096) = 4096 read(3, "s.')\nflags.DEFINE_integer('save"..., 4096) = 4096 read(3, "pointNotFoundException(Exception"..., 4096) = 4096 read(3, "params.eval_during_training_at_s"..., 4096) = 4096 read(3, "filename, ''\n as_text = fil"..., 4096) = 4096 read(3, "\n (name,"..., 4096) = 4096 read(3, "loat(piece))\n except ValueE"..., 4096) = 4096 read(3, "tf.train.RMSPropOptimizer(\n "..., 4096) = 4096 read(3, "rning_rate_decay_factor)):\n "..., 4096) = 4096 read(3, "d not self.params.datasets_use_p"..., 4096) = 4096 read(3, ".sync_queue_devices = [self.para"..., 4096) = 4096 read(3, "eter_server'):\n raise Value"..., 4096) = 4096 read(3, "onstants.BenchmarkMode.TRAIN_AND"..., 4096) = 4096 read(3, "d batches per prepocessing group"..., 4096) = 4096 read(3, "ame: %s' % self.params.job_name)"..., 4096) = 4096 write(2, " ", 4 ) = 4 write(2, "return self._benchmark_train()\n", 31return self.benchmark_train() ) = 31 close(3) = 0 write(2, " File "/home/tf/benchma"..., 112 File "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", line 2085, in benchmark_train ) = 112 openat(AT_FDCWD, "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=164060, ...}) = 0 read(3, "# Copyright 2017 The TensorFlow "..., 4096) = 4096 read(3, "\n# the forward-only option, wh"..., 4096) = 4096 read(3, " 'The autotune thresh"..., 4096) = 4096 read(3, " 'the number of s"..., 4096) = 4096 read(3, "ed, after the graph has been par"..., 4096) = 4096 read(3, " maximum number of GPU Ops that "..., 4096) = 4096 read(3, "or the MultiDeviceIterator that "..., 4096) = 4096 read(3, "INE_integer('kmp_settings', 1,\n "..., 4096) = 4096 read(3, " 'collective_all_redu"..., 4096) = 4096 read(3, "s.')\nflags.DEFINE_integer('save"..., 4096) = 4096 read(3, "pointNotFoundException(Exception"..., 4096) = 4096 read(3, "params.eval_during_training_at_s"..., 4096) = 4096 read(3, "filename, ''\n as_text = fil"..., 4096) = 4096 read(3, "\n (name,"..., 4096) = 4096 read(3, "loat(piece))\n except ValueE"..., 4096) = 4096 read(3, "tf.train.RMSPropOptimizer(\n "..., 4096) = 4096 read(3, "rning_rate_decay_factor)):\n "..., 4096) = 4096 read(3, "d not self.params.datasets_use_p"..., 4096) = 4096 read(3, ".sync_queue_devices = [self.para"..., 4096) = 4096 read(3, "eter_server'):\n raise Value"..., 4096) = 4096 read(3, "onstants.BenchmarkMode.TRAIN_AND"..., 4096) = 4096 read(3, "d batches per prepocessing group"..., 4096) = 4096 read(3, "ame: %s' % self.params.job_name)"..., 4096) = 4096 read(3, "not None:\n # We might rei"..., 4096) = 4096 read(3, "f.batch_size)\n mlperf.logge"..., 4096) = 4096 write(2, " ", 4 ) = 4 write(2, "return self.benchmark_graph(res"..., 70return self.benchmark_graph(result_to_benchmark, eval_build_results) ) = 70 close(3) = 0 write(2, " File "/home/tf/benchma"..., 112 File "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", line 2294, in benchmark_graph ) = 112 openat(AT_FDCWD, "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=164060, ...}) = 0 read(3, "# Copyright 2017 The TensorFlow "..., 4096) = 4096 read(3, "\n# the forward-only option, wh"..., 4096) = 4096 read(3, " 'The autotune thresh"..., 4096) = 4096 read(3, " 'the number of s"..., 4096) = 4096 read(3, "ed, after the graph has been par"..., 4096) = 4096 read(3, " maximum number of GPU Ops that "..., 4096) = 4096 read(3, "or the MultiDeviceIterator that "..., 4096) = 4096 read(3, "INE_integer('kmp_settings', 1,\n "..., 4096) = 4096 read(3, " 'collective_all_redu"..., 4096) = 4096 read(3, "s.')\nflags.DEFINE_integer('save"..., 4096) = 4096 read(3, "pointNotFoundException(Exception"..., 4096) = 4096 read(3, "params.eval_during_training_at_s"..., 4096) = 4096 read(3, "filename, ''\n as_text = fil"..., 4096) = 4096 read(3, "\n (name,"..., 4096) = 4096 read(3, "loat(piece))\n except ValueE"..., 4096) = 4096 read(3, "tf.train.RMSPropOptimizer(\n "..., 4096) = 4096 read(3, "rning_rate_decay_factor)):\n "..., 4096) = 4096 read(3, "d not self.params.datasets_use_p"..., 4096) = 4096 read(3, ".sync_queue_devices = [self.para"..., 4096) = 4096 read(3, "eter_server'):\n raise Value"..., 4096) = 4096 read(3, "onstants.BenchmarkMode.TRAIN_AND"..., 4096) = 4096 read(3, "d batches per prepocessing group"..., 4096) = 4096 read(3, "ame: %s' % self.params.job_name)"..., 4096) = 4096 read(3, "not None:\n # We might rei"..., 4096) = 4096 read(3, "f.batch_size)\n mlperf.logge"..., 4096) = 4096 read(3, "ributed_collective) and\n "..., 4096) = 4096 read(3, ".broadcast_global_variables(0)\n "..., 4096) = 4096 write(2, " ", 4 ) = 4 write(2, "is_chief, summary_writer, profil"..., 36is_chief, summary_writer, profiler) ) = 36 close(3) = 0 write(2, " File "/home/tf/benchma"..., 118 File "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", line 2430, in benchmark_with_session ) = 118 openat(AT_FDCWD, "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=164060, ...}) = 0 read(3, "# Copyright 2017 The TensorFlow "..., 4096) = 4096 read(3, "\n# the forward-only option, wh"..., 4096) = 4096 read(3, " 'The autotune thresh"..., 4096) = 4096 read(3, " 'the number of s"..., 4096) = 4096 read(3, "ed, after the graph has been par"..., 4096) = 4096 read(3, " maximum number of GPU Ops that "..., 4096) = 4096 read(3, "or the MultiDeviceIterator that "..., 4096) = 4096 read(3, "INE_integer('kmp_settings', 1,\n "..., 4096) = 4096 read(3, " 'collective_all_redu"..., 4096) = 4096 read(3, "s.')\nflags.DEFINE_integer('save"..., 4096) = 4096 read(3, "pointNotFoundException(Exception"..., 4096) = 4096 read(3, "params.eval_during_training_at_s"..., 4096) = 4096 read(3, "filename, ''\n as_text = fil"..., 4096) = 4096 read(3, "\n (name,"..., 4096) = 4096 read(3, "loat(piece))\n except ValueE"..., 4096) = 4096 read(3, "tf.train.RMSPropOptimizer(\n "..., 4096) = 4096 read(3, "rning_rate_decay_factor)):\n "..., 4096) = 4096 read(3, "d not self.params.datasets_use_p"..., 4096) = 4096 read(3, ".sync_queue_devices = [self.para"..., 4096) = 4096 read(3, "eter_server'):\n raise Value"..., 4096) = 4096 read(3, "onstants.BenchmarkMode.TRAIN_AND"..., 4096) = 4096 read(3, "d batches per prepocessing group"..., 4096) = 4096 read(3, "ame: %s' % self.params.job_name)"..., 4096) = 4096 read(3, "not None:\n # We might rei"..., 4096) = 4096 read(3, "f.batch_size)\n mlperf.logge"..., 4096) = 4096 read(3, "ributed_collective) and\n "..., 4096) = 4096 read(3, ".broadcast_global_variables(0)\n "..., 4096) = 4096 read(3, " named tensors/ops list, fe"..., 4096) = 4096 read(3, "fo.execution_barrier:\n "..., 4096) = 4096 write(2, " ", 4 ) = 4 write(2, "collective_graph_key=collective"..., 43collective_graph_key=collective_graph_key) ) = 43 close(3) = 0 write(2, " File "/home/tf/benchma"..., 113 File "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", line 869, in benchmark_one_step ) = 113 openat(AT_FDCWD, "/home/tf/benchmarks/scripts/tf_cnn_benchmarks/benchmark_cnn.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=164060, ...}) = 0 read(3, "# Copyright 2017 The TensorFlow "..., 4096) = 4096 read(3, "\n# the forward-only option, wh"..., 4096) = 4096 read(3, " 'The autotune thresh"..., 4096) = 4096 read(3, " 'the number of s"..., 4096) = 4096 read(3, "ed, after the graph has been par"..., 4096) = 4096 read(3, " maximum number of GPU Ops that "..., 4096) = 4096 read(3, "or the MultiDeviceIterator that "..., 4096) = 4096 read(3, "INE_integer('kmp_settings', 1,\n "..., 4096) = 4096 read(3, " 'collective_all_redu"..., 4096) = 4096 read(3, "s.')\nflags.DEFINE_integer('save"..., 4096) = 4096 read(3, "pointNotFoundException(Exception"..., 4096) = 4096 read(3, "params.eval_during_training_at_s"..., 4096) = 4096 write(2, " ", 4 ) = 4 write(2, "results = sess.run(fetches, opti"..., 76results = sess.run(fetches, options=run_options, run_metadata=run_metadata) ) = 76 close(3) = 0 write(2, " File "/root/.env/tf1.15/lib/py"..., 114 File "/root/.env/tf1.15/lib/python2.7/site-packages/tensorflow_core/python/client/session.py", line 956, in run ) = 114 openat(AT_FDCWD, "/root/.env/tf1.15/lib/python2.7/site-packages/tensorflow_core/python/client/session.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=69365, ...}) = 0 read(3, "# Copyright 2015 The TensorFlow "..., 4096) = 4096 read(3, "d to feed a run, feed_fn2 to s"..., 4096) = 4096 read(3, ", tensor_type):\n raise Valu"..., 4096) = 4096 read(3, ": Callable as returned by a fetc"..., 4096) = 4096 read(3, "ass.\n """\n values = get_a"..., 4096) = 4096 read(3, "\n i = 0\n j = 0\n for is"..., 4096) = 4096 read(3, "dles = []\n\n if config is None"..., 4096) = 4096 read(3, "\n\n @Property\n def sess_str(sel"..., 4096) = 4096 read(3, "he value of tensors in the graph"..., 4096) = 4096 write(2, " ", 4 ) = 4 write(2, "run_metadata_ptr)\n", 18run_metadata_ptr) ) = 18 close(3) = 0 write(2, " File "/root/.env/tf1.15/lib/py"..., 116 File "/root/.env/tf1.15/lib/python2.7/site-packages/tensorflow_core/python/client/session.py", line 1180, in _run ) = 116 openat(AT_FDCWD, "/root/.env/tf1.15/lib/python2.7/site-packages/tensorflow_core/python/client/session.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=69365, ...}) = 0 read(3, "# Copyright 2015 The TensorFlow "..., 4096) = 4096 read(3, "d to feed a run, feed_fn2 to s"..., 4096) = 4096 read(3, ", tensor_type):\n raise Valu"..., 4096) = 4096 read(3, ": Callable as returned by a fetc"..., 4096) = 4096 read(3, "ass.\n """\n values = get_a"..., 4096) = 4096 read(3, "\n i = 0\n j = 0\n for is"..., 4096) = 4096 read(3, "dles = []\n\n if config is None"..., 4096) = 4096 read(3, "\n\n @Property\n def sess_str(sel"..., 4096) = 4096 read(3, "he value of tensors in the graph"..., 4096) = 4096 read(3, "tionary\n whose values are"..., 4096) = 4096 read(3, "d Session.')\n if self.graph.v"..., 4096) = 4096 write(2, " ", 4 ) = 4 write(2, "feed_dict_tensor, options, run_m"..., 41feed_dict_tensor, options, run_metadata) ) = 41 close(3) = 0 write(2, " File "/root/.env/tf1.15/lib/py"..., 119 File "/root/.env/tf1.15/lib/python2.7/site-packages/tensorflow_core/python/client/session.py", line 1359, in _do_run ) = 119 openat(AT_FDCWD, "/root/.env/tf1.15/lib/python2.7/site-packages/tensorflow_core/python/client/session.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=69365, ...}) = 0 read(3, "# Copyright 2015 The TensorFlow "..., 4096) = 4096 read(3, "d to feed a run, feed_fn2 to s"..., 4096) = 4096 read(3, ", tensor_type):\n raise Valu"..., 4096) = 4096 read(3, ": Callable as returned by a fetc"..., 4096) = 4096 read(3, "ass.\n """\n values = get_a"..., 4096) = 4096 read(3, "\n i = 0\n j = 0\n for is"..., 4096) = 4096 read(3, "dles = []\n\n if config is None"..., 4096) = 4096 read(3, "\n\n @Property\n def sess_str(sel"..., 4096) = 4096 read(3, "he value of tensors in the graph"..., 4096) = 4096 read(3, "tionary\n whose values are"..., 4096) = 4096 read(3, "d Session.')\n if self.graph.v"..., 4096) = 4096 read(3, "d_list)arguments whose types\n "..., 4096) = 4096 read(3, "session.TF_DeleteBuffer(run_meta"..., 4096) = 4096 write(2, " ", 4 ) = 4 write(2, "run_metadata)\n", 14run_metadata) ) = 14 close(3) = 0 write(2, " File \"/root/.env/tf1.15/lib/py"..., 120 File "/root/.env/tf1.15/lib/python2.7/site-packages/tensorflow_core/python/client/session.py", line 1384, in _do_call ) = 120 openat(AT_FDCWD, "/root/.env/tf1.15/lib/python2.7/site-packages/tensorflow_core/python/client/session.py", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=69365, ...}) = 0 read(3, "# Copyright 2015 The TensorFlow "..., 4096) = 4096 read(3, "d to feed a run,feed_fn2 to s"..., 4096) = 4096 read(3, ", tensor_type):\n raise Valu"..., 4096) = 4096 read(3, ": Callable as returned by a fetc"..., 4096) = 4096 read(3, "ass.\n \"\"\"\n values = _geta"..., 4096) = 4096 read(3, "\n i = 0\n j = 0\n for is"..., 4096) = 4096 read(3, "dles = []\n\n if config is None"..., 4096) = 4096 read(3, "\n\n @property\n def sess_str(sel"..., 4096) = 4096 read(3, "he value of tensors in the graph"..., 4096) = 4096 read(3, "tionary\n whose values are"..., 4096) = 4096 read(3, "d Session.')\n if self.graph.v"..., 4096) = 4096 read(3, "d_list) arguments whose types\n "..., 4096) = 4096 read(3, "session.TF_DeleteBuffer(run_meta"..., 4096) = 4096 read(3, "ssage, self._graph)\n if 'on"..., 4096) = 4096 write(2, " ", 4 ) = 4 write(2, "raise type(e)(node_def, op, mess"..., 37raise type(e)(node_def, op, message) ) = 37 close(3) = 0 write(2, "tensorflow.python.framework.erro"..., 39tensorflow.python.framework.errors_impl) = 39 write(2, ".", 1.) = 1 write(2, "InvalidArgumentError", 20InvalidArgumentError) = 20 write(2, ": ", 2: ) = 2 write(2, "Retval[0] does not have value", 29Retval[0] does not have value) = 29 write(2, "\n", 1 ) = 1 unlinkat(AT_FDCWD, "/tmp/tmp5HeJNi.py", 0) = 0 unlinkat(AT_FDCWD, "/tmp/tmpuWUoTb.py", 0) = 0 unlinkat(AT_FDCWD, "/tmp/tmpRaOsTY.py", 0) = 0 unlinkat(AT_FDCWD, "/tmp/tmpt0VBwu.py", 0) = 0 unlinkat(AT_FDCWD, "/tmp/tmpZMqsKy.py", 0) = 0 unlinkat(AT_FDCWD, "/tmp/tmpl2wn88.py", 0) = 0 unlinkat(AT_FDCWD, "/tmp/tmpcDvUvG.py", 0) = 0 unlinkat(AT_FDCWD, "/tmp/tmpgPPLwh.py", 0) = 0 rt_sigaction(SIGINT, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=0}, {sa_handler=0xaaaada8a4260, sa_mask=[], sa_flags=0}, 8) = 0 munmap(0xfffd94779000, 262144) = 0 munmap(0xfffd8d1bc000, 262144) = 0 munmap(0xfffd8d33b000, 262144) = 0 munmap(0xfffd946f9000, 262144) = 0 munmap(0xfffdac6f9000, 262144) = 0 munmap(0xfffd8d2bb000, 262144) = 0 munmap(0xffff77141000, 262144) = 0 munmap(0xfffd8cefc000, 262144) = 0 munmap(0xfffda4339000, 262144) = 0 close(11) = 0 futex(0xffff8071a674, FUTEX_WAKE_PRIVATE, 2147483647) = 0 fstat(0, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 0), ...}) = 0 fstat(0, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 0), ...}) = 0 fstat(0, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 0), ...}) = 0 exit_group(1) = ? +++ exited with 1 +++

[Danliran]

I alse post this issue in tensorflow https://github.com/tensorflow/tensorflow/issues/40213

[reedwm]

Unfortunately, tf_cnn_benchmarks is no longer maintained so its unlikely it will get resolved. I recommend trying the TensorFlow Official Models instead. I'm unsure what the state of aarch64 in TensorFlow however, so am unsure if the Official Models will work either.

@gunan do you know how well supported aarch64 is in TensorFlow?

[gunan]

I wouldn't know. @petewarden may?