data folder for one_shot_learning.py ?
softgearko opened this issue · comments
I got an error when I run 'python one_shot_learning.py'. I think, it occurred because there is no data folder .
Could you explain how to get omniglot dataset and set up the 'data' folder?
Thank you in advanced.
$ python3 one_shot_learning.py
2018-01-22 17:12:28.831022: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
2018-01-22 17:12:28.972256: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:892] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2018-01-22 17:12:28.972609: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1030] Found device 0 with properties:
name: TITAN X (Pascal) major: 6 minor: 1 memoryClockRate(GHz): 1.531
pciBusID: 0000:01:00.0
totalMemory: 11.90GiB freeMemory: 11.76GiB
2018-01-22 17:12:28.972622: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1120] Creating TensorFlow device (/device:GPU:0) -> (device: 0, name: TITAN X (Pascal), pci bus id: 0000:01:00.0, compute capability: 6.1)
Namespace(augment=True, batch_size=16, debug=False, image_height=20, image_width=20, label_type='one_hot', learning_rate=0.001, memory_size=128, memory_vector_dim=40, mode='train', model='MANN', n_classes=5, n_test_classes=423, n_train_classes=1200, num_epoches=100000, output_dim=5, read_head_num=4, restore_training=False, rnn_num_layers=1, rnn_size=200, save_dir='./save/one_shot_learning', seq_length=50, shift_range=1, tensorboard_dir='./summary/one_shot_learning', test_batch_num=100, write_head_num=1)
1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th batch loss
Traceback (most recent call last):
File "one_shot_learning.py", line 156, in
main()
File "one_shot_learning.py", line 38, in main
train(args)
File "one_shot_learning.py", line 71, in train
label_type=args.label_type)
File "/mnt/data1/softgear/snowkylin_ntm_mann/ntm/utils.py", line 66, in fetch_batch
classes = [np.random.choice(range(len(data)), replace=False, size=n_classes) for _ in range(batch_size)]
File "/mnt/data1/softgear/snowkylin_ntm_mann/ntm/utils.py", line 66, in
classes = [np.random.choice(range(len(data)), replace=False, size=n_classes) for _ in range(batch_size)]
File "mtrand.pyx", line 1121, in mtrand.RandomState.choice
ValueError: a must be non-empty
$