thuml / hashgan Goto Github PK

Hi, i have a problem about finding the input v.In the code it seems that there isnt the input s.and
i have not found the loss about the v and s.

您好，请问论文中关于相似嵌入的内容在代码的哪部分体现了？非常感谢

Where is the encoder implemented

Hi, thanks for sharing your code. It is a great job. But I meet some errors when I try this script.

ValueError Traceback (most recent call last)
in ()
422
423 mAP_ = util.MAPs(MAP_R)
--> 424 mAP_val = mAP_.get_mAPs_by_feature(db, test)
425 lib.plot.plot("mAP_feature", mAP_val)
426

/home/zoe/deeplearning/HashGAN/util.pyc in get_mAPs_by_feature(self, database, query)
135 rel = np.sum(imatch)
136 Lx = np.cumsum(imatch)
--> 137 Px = Lx.astype(float) / np.arange(1, self.R + 1, 1)
138 if rel != 0:
139 APx.append(np.sum(Px * imatch) / rel)

ValueError: operands could not be broadcast together with shapes (53952,) (54000,)

So what should I do to get this net worked？Thanks！

Hi! I am looking to do just inference and would love to avoid retraining on something like ImageNet (I don't have enough gpus). Would anyone be willing to provide a pretrained model that was done something like on ImageNet or NUS-WIDE or similar.

I would really appreciate it!! Thanks.

hello!I am interested in HashGAN and want to do a comparative experiment on coco and imagenet. Can you give me the configuration(yaml) on these two datasets?

Hello, thank you for sharing your code! Could please help me in solving this problem?

/home/umair/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:516: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint8 = np.dtype([("qint8", np.int8, 1)])
/home/umair/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:517: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint8 = np.dtype([("quint8", np.uint8, 1)])
/home/umair/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:518: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint16 = np.dtype([("qint16", np.int16, 1)])
/home/umair/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:519: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint16 = np.dtype([("quint16", np.uint16, 1)])
/home/umair/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:520: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint32 = np.dtype([("qint32", np.int32, 1)])
/home/umair/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:525: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
np_resource = np.dtype([("resource", np.ubyte, 1)])
/home/umair/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint8 = np.dtype([("qint8", np.int8, 1)])
/home/umair/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint8 = np.dtype([("quint8", np.uint8, 1)])
/home/umair/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint16 = np.dtype([("qint16", np.int16, 1)])
/home/umair/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint16 = np.dtype([("quint16", np.uint16, 1)])
/home/umair/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint32 = np.dtype([("qint32", np.int32, 1)])
/home/umair/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
np_resource = np.dtype([("resource", np.ubyte, 1)])
{'DATA': {'DATA_ROOT': './data/cifar10',
'DB_SIZE': 54000,
'IMAGE_DIR': './output/cifar10_step_1_ACGAN_SCALE_G_10.0/images',
'LABEL_DIM': 10,
'LIST_ROOT': './data_list/cifar10',
'LOG_DIR': './output/cifar10_step_1_ACGAN_SCALE_G_10.0/logs',
'MAP_R': 54000,
'MODEL_DIR': './output/cifar10_step_1_ACGAN_SCALE_G_10.0/models',
'OUTPUT_DIM': 3072,
'OUTPUT_DIR': './output/cifar10_step_1_ACGAN_SCALE_G_10.0',
'TEST_SIZE': 1000,
'USE_DATASET': 'cifar10',
'WIDTH_HEIGHT': 32},
'MODEL': {'ALEXNET_PRETRAINED_MODEL_PATH': './pretrained_models/reference_pretrain.npy',
'DIM': 64,
'DIM_D': 128,
'DIM_G': 128,
'D_ARCHITECTURE': 'NORM',
'D_PRETRAINED_MODEL_PATH': '',
'G_ARCHITECTURE': 'NORM',
'G_PRETRAINED_MODEL_PATH': '',
'HASH_DIM': 64},
'TRAIN': {'ACGAN_SCALE': 1.0,
'ACGAN_SCALE_FAKE': 1.0,
'ACGAN_SCALE_G': 10.0,
'BATCH_SIZE': 64,
'CHECKPOINT_FREQUENCY': 1000,
'CROSS_ENTROPY_ALPHA': 10,
'DECAY': True,
'EVALUATE_MODE': False,
'EVAL_FREQUENCY': 10000,
'FAKE_RATIO': 1.0,
'G_LR': 0.0001,
'ITERS': 100000,
'LR': 0.0001,
'NORMED_CROSS_ENTROPY': True,
'N_CRITIC': 5,
'SAMPLE_FREQUENCY': 1000,
'WGAN_SCALE': 1.0,
'WGAN_SCALE_G': 1.0,
'WGAN_SCALE_GP': 10.0}}
WARNING:tensorflow:From main.py:40: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.

WARNING:tensorflow:From main.py:147: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.

WARNING:tensorflow:From /home/umair/HashGAN/lib/architecture.py:109: The name tf.random_normal is deprecated. Please use tf.random.normal instead.

WARNING:tensorflow:From /home/umair/HashGAN/lib/architecture.py:43: The name tf.depth_to_space is deprecated. Please use tf.compat.v1.depth_to_space instead.

WARNING:tensorflow:From /home/umair/HashGAN/lib/ops.py:125: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.

WARNING:tensorflow:From /home/umair/HashGAN/lib/criterion.py:26: div (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Deprecated in favor of operator or tf.math.divide.
WARNING:tensorflow:From /home/umair/HashGAN/lib/criterion.py:41: The name tf.log is deprecated. Please use tf.math.log instead.

WARNING:tensorflow:From main.py:72: The name tf.summary.scalar is deprecated. Please use tf.compat.v1.summary.scalar instead.

WARNING:tensorflow:From main.py:97: The name tf.train.AdamOptimizer is deprecated. Please use tf.compat.v1.train.AdamOptimizer instead.

WARNING:tensorflow:From /home/umair/.local/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1205: add_dispatch_support..wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.where in 2.0, which has the same broadcast rule as np.where
WARNING:tensorflow:From main.py:100: The name tf.summary.merge is deprecated. Please use tf.compat.v1.summary.merge instead.

2019-09-30 15:36:36.266052: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2019-09-30 15:36:36.286147: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 4200000000 Hz
2019-09-30 15:36:36.286436: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x560d35ac3540 executing computations on platform Host. Devices:
2019-09-30 15:36:36.286451: I tensorflow/compiler/xla/service/service.cc:175] StreamExecutor device (0): ,
WARNING:tensorflow:From main.py:176: The name tf.summary.FileWriter is deprecated. Please use tf.compat.v1.summary.FileWriter instead.

computing param size
G Params:
generator.Input.W:0 (256,2048)
generator.Input.b:0 (2048)
generator.1.Shortcut.Filters:0 (1,1,128,128)
generator.1.Shortcut.Biases:0 (128)
generator.1.N1.offset:0 (128)
generator.1.N1.scale:0 (128)
generator.1.N1.moving_mean:0 (128) [no grad!]
generator.1.N1.moving_variance:0 (128) [no grad!]
generator.1.Conv1.Filters:0 (3,3,128,128)
generator.1.Conv1.Biases:0 (128)
generator.1.N2.offset:0 (128)
generator.1.N2.scale:0 (128)
generator.1.N2.moving_mean:0 (128) [no grad!]
generator.1.N2.moving_variance:0 (128) [no grad!]
generator.1.Conv2.Filters:0 (3,3,128,128)
generator.1.Conv2.Biases:0 (128)
generator.2.Shortcut.Filters:0 (1,1,128,128)
generator.2.Shortcut.Biases:0 (128)
generator.2.N1.offset:0 (128)
generator.2.N1.scale:0 (128)
generator.2.N1.moving_mean:0 (128) [no grad!]
generator.2.N1.moving_variance:0 (128) [no grad!]
generator.2.Conv1.Filters:0 (3,3,128,128)
generator.2.Conv1.Biases:0 (128)
generator.2.N2.offset:0 (128)
generator.2.N2.scale:0 (128)
generator.2.N2.moving_mean:0 (128) [no grad!]
generator.2.N2.moving_variance:0 (128) [no grad!]
generator.2.Conv2.Filters:0 (3,3,128,128)
generator.2.Conv2.Biases:0 (128)
generator.3.Shortcut.Filters:0 (1,1,128,128)
generator.3.Shortcut.Biases:0 (128)
generator.3.N1.offset:0 (128)
generator.3.N1.scale:0 (128)
generator.3.N1.moving_mean:0 (128) [no grad!]
generator.3.N1.moving_variance:0 (128) [no grad!]
generator.3.Conv1.Filters:0 (3,3,128,128)
generator.3.Conv1.Biases:0 (128)
generator.3.N2.offset:0 (128)
generator.3.N2.scale:0 (128)
generator.3.N2.moving_mean:0 (128) [no grad!]
generator.3.N2.moving_variance:0 (128) [no grad!]
generator.3.Conv2.Filters:0 (3,3,128,128)
generator.3.Conv2.Biases:0 (128)
generator.OutputN.offset:0 (128)
generator.OutputN.scale:0 (128)
generator.OutputN.moving_mean:0 (128) [no grad!]
generator.OutputN.moving_variance:0 (128) [no grad!]
generator.Output.Filters:0 (3,3,128,3)
generator.Output.Biases:0 (3)
Total param count: 1,468,419
D Params:
discriminator.1.Shortcut.Filters:0 (1,1,3,128)
discriminator.1.Shortcut.Biases:0 (128)
discriminator.1.Conv1.Filters:0 (3,3,3,128)
discriminator.1.Conv1.Biases:0 (128)
discriminator.1.Conv2.Filters:0 (3,3,128,128)
discriminator.1.Conv2.Biases:0 (128)
discriminator.2.Shortcut.Filters:0 (1,1,128,128)
discriminator.2.Shortcut.Biases:0 (128)
discriminator.2.Conv1.Filters:0 (3,3,128,128)
discriminator.2.Conv1.Biases:0 (128)
discriminator.2.Conv2.Filters:0 (3,3,128,128)
discriminator.2.Conv2.Biases:0 (128)
discriminator.3.Conv1.Filters:0 (3,3,128,128)
discriminator.3.Conv1.Biases:0 (128)
discriminator.3.Conv2.Filters:0 (3,3,128,128)
discriminator.3.Conv2.Biases:0 (128)
discriminator.4.Conv1.Filters:0 (3,3,128,128)
discriminator.4.Conv1.Biases:0 (128)
discriminator.4.Conv2.Filters:0 (3,3,128,128)
discriminator.4.Conv2.Biases:0 (128)
discriminator.Output.W:0 (128,1)
discriminator.Output.b:0 (1)
discriminator.ACGANOutput.W:0 (128,64)
discriminator.ACGANOutput.b:0 (64)
Total param count: 1,062,081
initializing global variables
2019-09-30 15:36:37.062999: W tensorflow/compiler/jit/mark_for_compilation_pass.cc:1412] (One-time warning): Not using XLA:CPU for cluster because envvar TF_XLA_FLAGS=--tf_xla_cpu_global_jit was not set. If you want XLA:CPU, either set that envvar, or use experimental_jit_scope to enable XLA:CPU. To confirm that XLA is active, pass --vmodule=xla_compilation_cache=1 (as a proper command-line flag, not via TF_XLA_FLAGS) or set the envvar XLA_FLAGS=--xla_hlo_profile.
WARNING:tensorflow:From main.py:187: The name tf.train.Saver is deprecated. Please use tf.compat.v1.train.Saver instead.

training
Training: 0%| | 0/100000 [00:00<?, ?it/s]cannot open train/1_3458.jpg 0 0 0 1 0 0 0 0 0 0

cannot open test/999_9471.jpg 0 0 0 0 0 0 1 0 0 0

cannot open train/4_1709.jpg 0 0 0 0 0 0 1 0 0 0

cannot open train/4_8781.jpg 0 0 0 0 1 0 0 0 0 0

cannot open test/999_2.jpg 0 0 0 0 0 0 0 0 1 0

cannot open train/4_6735.jpg 0 0 0 0 0 0 0 0 1 0

cannot open train/3_238.jpg 0 1 0 0 0 0 0 0 0 0

cannot open train/1_2437.jpg 0 0 0 0 0 0 0 1 0 0

cannot open test/999_2761.jpg 0 0 0 0 0 0 0 1 0 0

cannot open train/0_9239.jpg 0 0 0 0 0 0 0 1 0 0

cannot open test/999_7562.jpg 0 0 1 0 0 0 0 0 0 0

cannot open test/999_6229.jpg 0 0 0 0 0 0 1 0 0 0

cannot open train/2_1646.jpg 0 0 0 0 0 0 0 0 1 0

cannot open train/3_1760.jpg 0 0 0 0 0 0 0 0 0 1

cannot open train/0_185.jpg 1 0 0 0 0 0 0 0 0 0

cannot open train/0_7860.jpg 0 0 0 0 0 0 1 0 0 0

cannot open train/3_1497.jpg 0 0 0 0 0 0 0 0 1 0

cannot open test/999_9007.jpg 0 0 1 0 0 0 0 0 0 0

cannot open train/0_4707.jpg 0 0 0 0 1 0 0 0 0 0

cannot open train/4_6205.jpg 0 0 0 1 0 0 0 0 0 0

cannot open train/2_7457.jpg 0 0 0 0 0 0 0 0 0 1

cannot open train/0_4004.jpg 0 0 0 0 0 1 0 0 0 0

cannot open train/2_5493.jpg 0 0 0 0 0 0 1 0 0 0

cannot open train/1_6988.jpg 0 0 0 0 0 0 0 0 1 0

cannot open train/0_191.jpg 0 0 0 0 0 0 0 1 0 0

cannot open train/2_744.jpg 0 0 0 0 0 0 0 1 0 0

cannot open test/999_538.jpg 0 0 0 0 0 0 0 0 0 1

cannot open train/4_3470.jpg 0 0 0 0 1 0 0 0 0 0

cannot open train/3_994.jpg 0 0 0 0 0 1 0 0 0 0

cannot open train/2_7388.jpg 0 0 0 0 1 0 0 0 0 0

cannot open train/3_436.jpg 0 0 0 0 1 0 0 0 0 0

cannot open train/1_4124.jpg 1 0 0 0 0 0 0 0 0 0

cannot open train/1_1148.jpg 0 0 0 0 0 0 0 0 1 0

cannot open train/0_1428.jpg 1 0 0 0 0 0 0 0 0 0

cannot open test/999_6256.jpg 0 0 0 0 0 0 0 0 0 1

cannot open train/4_489.jpg 0 0 0 0 0 1 0 0 0 0

cannot open train/2_2163.jpg 0 1 0 0 0 0 0 0 0 0

cannot open train/0_7556.jpg 0 0 0 0 0 0 0 0 0 1

cannot open test/999_2245.jpg 0 0 0 0 0 0 0 0 1 0

cannot open test/999_9847.jpg 1 0 0 0 0 0 0 0 0 0

cannot open train/4_523.jpg 0 0 0 0 1 0 0 0 0 0

cannot open train/1_7046.jpg 0 0 0 1 0 0 0 0 0 0

cannot open train/2_9469.jpg 0 1 0 0 0 0 0 0 0 0

cannot open test/999_5002.jpg 0 0 0 0 0 0 0 0 1 0

cannot open train/2_1446.jpg 0 0 1 0 0 0 0 0 0 0

cannot open train/2_7428.jpg 0 0 0 0 1 0 0 0 0 0

cannot open train/2_3093.jpg 0 0 0 0 0 1 0 0 0 0

cannot open train/3_5260.jpg 0 0 0 0 0 0 0 0 0 1

cannot open train/2_7730.jpg 0 1 0 0 0 0 0 0 0 0

cannot open train/1_1671.jpg 1 0 0 0 0 0 0 0 0 0

cannot open train/2_4955.jpg 0 0 0 0 0 0 1 0 0 0

cannot open test/999_8755.jpg 0 0 0 0 0 0 1 0 0 0

cannot open train/2_7013.jpg 0 0 0 1 0 0 0 0 0 0

cannot open train/2_7632.jpg 0 0 0 0 0 0 1 0 0 0

cannot open train/4_5532.jpg 0 0 1 0 0 0 0 0 0 0

cannot open test/999_1762.jpg 0 0 0 0 1 0 0 0 0 0

cannot open train/1_7367.jpg 0 0 0 0 0 0 0 0 0 1

cannot open train/4_9735.jpg 1 0 0 0 0 0 0 0 0 0

cannot open train/1_1554.jpg 1 0 0 0 0 0 0 0 0 0

cannot open test/999_3533.jpg 0 0 1 0 0 0 0 0 0 0

cannot open test/999_7383.jpg 0 0 0 0 1 0 0 0 0 0

cannot open train/1_8331.jpg 0 1 0 0 0 0 0 0 0 0

cannot open train/4_5356.jpg 0 0 0 0 0 0 0 0 0 1

cannot open test/999_3276.jpg 0 0 0 0 1 0 0 0 0 0

Training: 0%| | 0/100000 [00:00<?, ?it/s]
Traceback (most recent call last):
File "main.py", line 270, in
main(config)
File "main.py", line 224, in main
summary_disc, _ = session.run([model.summary_disc, model.train_op_disc], feed_dict=get_feed_dict())
File "main.py", line 207, in get_feed_dict
labeled_data, labeled_labels = gen()
File "/home/umair/HashGAN/lib/dataloader.py", line 138, in generator
for images_iter_, labels_iter_ in gen():
File "/home/umair/HashGAN/lib/dataloader.py", line 110, in get_epoch
data = np.transpose(data, (0, 3, 1, 2))
File "<array_function internals>", line 6, in transpose
File "/home/umair/.local/lib/python3.6/site-packages/numpy/core/fromnumeric.py", line 650, in transpose
return _wrapfunc(a, 'transpose', axes)
File "/home/umair/.local/lib/python3.6/site-packages/numpy/core/fromnumeric.py", line 61, in _wrapfunc
return bound(*args, **kwds)
ValueError: axes don't match array

Hello, I'd like to ask you about the comparison of P-R curve with other methods in the paper. Whether other methods should be implemented once or whether there are existing CSV files

You can download the pretrained Generator models in the release page and modify config file to use the pretrained models.

Hi ,I don't understand how to do this step.

cifar10_G_99999.ckpt.zip
Source code
(zip)

Source code
(tar.gz)
which one?