Google Colaboratory上でのFatal Python error: Segmentation fault

前提・実現したいこと

Google Colaboratoryを利用して無料GPUでDeepLabを動作させる
こちらを参考にDeepLabv3+の実装を行っています．
モデルのテストに成功し，PASCAL VOC 2012によるテストを行おうとしたところ
Fatal Python error: Segmentation fault
が発生しました．

発生している問題・エラーメッセージ

<多くのWARNINGメッセージ(deprecated)>
...
INFO:tensorflow:Initializing model from path: datasets/pascal_voc_seg/init_models/deeplabv3_pascal_train_aug/model.ckpt
I0814 06:17:46.054309 139962412631936 train_utils.py:207] Initializing model from path: datasets/pascal_voc_seg/init_models/deeplabv3_pascal_train_aug/model.ckpt
WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow_core/contrib/slim/python/slim/learning.py:742: Supervisor.__init__ (from tensorflow.python.training.supervisor) is deprecated and will be removed in a future version.
Instructions for updating:
Please switch to tf.train.MonitoredTrainingSession
W0814 06:17:49.567995 139962412631936 deprecation.py:323] From /usr/local/lib/python3.6/dist-packages/tensorflow_core/contrib/slim/python/slim/learning.py:742: Supervisor.__init__ (from tensorflow.python.training.supervisor) is deprecated and will be removed in a future version.
Instructions for updating:
Please switch to tf.train.MonitoredTrainingSession
2020-08-14 06:17:50.492689: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2020-08-14 06:17:50.497361: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 2200000000 Hz
2020-08-14 06:17:50.497559: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x276319c0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2020-08-14 06:17:50.497589: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): Host, Default Version
2020-08-14 06:17:50.602823: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:983] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-08-14 06:17:50.603544: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x27631800 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:
2020-08-14 06:17:50.603574: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): Tesla T4, Compute Capability 7.5
2020-08-14 06:17:50.603750: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:983] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-08-14 06:17:50.604330: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1639] Found device 0 with properties: 
name: Tesla T4 major: 7 minor: 5 memoryClockRate(GHz): 1.59
pciBusID: 0000:00:04.0
2020-08-14 06:17:50.604434: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.0
2020-08-14 06:17:50.604459: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10.0
2020-08-14 06:17:50.604484: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcufft.so.10.0
2020-08-14 06:17:50.604505: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcurand.so.10.0
2020-08-14 06:17:50.604525: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusolver.so.10.0
2020-08-14 06:17:50.604544: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusparse.so.10.0
2020-08-14 06:17:50.604564: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7
2020-08-14 06:17:50.604639: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:983] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-08-14 06:17:50.605243: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:983] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-08-14 06:17:50.605759: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1767] Adding visible gpu devices: 0
2020-08-14 06:17:50.605857: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.0
2020-08-14 06:17:50.607042: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1180] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-08-14 06:17:50.607070: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1186]      0 
2020-08-14 06:17:50.607081: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1199] 0:   N 
2020-08-14 06:17:50.607193: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:983] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-08-14 06:17:50.607739: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:983] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-08-14 06:17:50.608259: W tensorflow/core/common_runtime/gpu/gpu_bfc_allocator.cc:39] Overriding allow_growth setting because the TF_FORCE_GPU_ALLOW_GROWTH environment variable is set. Original config value was 0.
2020-08-14 06:17:50.608299: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1325] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 14221 MB memory) -> physical GPU (device: 0, name: Tesla T4, pci bus id: 0000:00:04.0, compute capability: 7.5)
INFO:tensorflow:Running local_init_op.
I0814 06:17:56.731822 139962412631936 session_manager.py:500] Running local_init_op.
INFO:tensorflow:Done running local_init_op.
I0814 06:17:57.016979 139962412631936 session_manager.py:502] Done running local_init_op.
INFO:tensorflow:Starting Session.
I0814 06:18:04.286493 139962412631936 learning.py:754] Starting Session.
INFO:tensorflow:Saving checkpoint to path datasets/pascal_voc_seg/exp/train_on_trainval_set/train/model.ckpt
I0814 06:18:04.310412 139957582690048 supervisor.py:1117] Saving checkpoint to path datasets/pascal_voc_seg/exp/train_on_trainval_set/train/model.ckpt
INFO:tensorflow:Starting Queues.
I0814 06:18:04.310543 139962412631936 learning.py:768] Starting Queues.
Fatal Python error: Segmentation fault

該当のソースコード

python
1%time !python train.py \
2  --logtostderr \
3  --train_split="trainval" \
4  --model_variant="xception_65" \
5  --atrous_rates=6 \
6  --atrous_rates=12 \
7  --atrous_rates=18 \
8  --output_stride=16 \
9  --decoder_output_stride=4 \
10  --train_crop_size="513,513" \
11  --train_batch_size=4 \
12  --training_number_of_steps=$NUM_ITERATIONS \
13  --fine_tune_batch_norm=true \
14  --tf_initial_checkpoint=$INIT_FOLDER/deeplabv3_pascal_train_aug/model.ckpt \
15  --train_logdir=$TRAIN_LOGDIR \
16  --dataset_dir=$PASCAL_DATASET