InvalidArgumentError : restore the weight of Inception_v4

[zacharyclam]

System information

• OS Platform and Distribution : windows10
• TensorFlow installed from : Source
• TensorFlow version : 1.7.0
• Python version: Python 3.6.3
• Bazel version: 0.11.1
• GCC/Compiler version : 5.4.0
• CUDA/cuDNN version: no
• GPU model and memory: no
• Exact command to reproduce:

Describe the problem

I'm training a deep neural network.I want to restore the weight of Inception_v4 ,but there are some errors.I think the define of network is right but I can't restore the weight of model.Is it a bug of different version.

The errors:

InvalidArgumentError (see above for traceback): Assign requires shapes of both tensors to match. lhs shape= [3,3,128,768] rhs shape= 5,5,128,768, use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/device:CPU:0"](InceptionV4/AuxLogits/Conv2d_2a/weights, save/RestoreV2:9)]]

Source code

import scipy.misc as misc
import pickle
import tensorflow as tf
from tqdm import tqdm
import numpy as np
import argparse
import fnmatch
import sys
import os

from nets.inception_v4 import *
sys.path.insert(0, 'nets/')
slim = tf.contrib.slim

def getPaths(data_dir):
    image_paths = []
    # add more extensions if need be
    ps = ['jpg', 'jpeg', 'JPG', 'JPEG', 'bmp', 'BMP', 'png', 'PNG']
    for p in ps:
        pattern = '*.' + p
        for d, s, fList in os.walk(data_dir):
            for filename in fList:
                if fnmatch.fnmatch(filename, pattern):
                    fname_ = os.path.join(d, filename)
                    image_paths.append(fname_)
    return image_paths

if name == 'main':
    parser = argparse.ArgumentParser()
    parser.add_argument('--data_dir', required=True, type=str, help='Directory images are in. Searches recursively.')
    parser.add_argument('--model', required=True, type=str, help='Model to use')
    parser.add_argument('--checkpoint_file', required=True, type=str, help='Model file')
    a = parser.parse_args()
    data_dir = a.data_dir
    model = a.model
    checkpoint_file = a.checkpoint_file
    # I only have these because I thought some take in size of (299,299), but maybe not
    height, width, channels = 224, 224, 3 
    x = tf.placeholder(tf.float32, shape=(1, height, width, channels))
    
    arg_scope = inception_v4_arg_scope()
    with slim.arg_scope(arg_scope):
        logits, end_points = inception_v4(x, is_training=False, num_classes=1001)
        features = end_points['PreLogitsFlatten']
     
    sess = tf.Session()
    saver = tf.train.Saver()
    saver.restore(sess, checkpoint_file)
    
    feat_dict = {}
    paths = getPaths(data_dir)
    print('Computing features...')
    for path in tqdm(paths):
        image = misc.imread(path)
        image = misc.imresize(image, (height, width))
        image = np.expand_dims(image, 0)
        feat = np.squeeze(sess.run(features, feed_dict={x: image}))
        feat_dict[path] = feat
    
    try:
        os.makedirs('features/')
    except:
        pass
    exp_pkl = open('features/' + model + '_features.pkl', 'wb')
    data = pickle.dumps(feat_dict)
    exp_pkl.write(data)
    exp_pkl.close()

Log：

D:\PythonProject\Compute-Features-WithModel>python compute_features.py --data_dir=jaffe/ --checkpoint_file=inception_v4.ckpt --model=inception_v4
Use the retry module or similar alternatives.
Traceback (most recent call last):
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\client\session.py", line 1327, in _do_call
    return fn(*args)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\client\session.py", line 1312, in _run_fn
    options, feed_dict, fetch_list, target_list, run_metadata)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\client\session.py", line 1420, in _call_tf_sessionrun
    status, run_metadata)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\framework\errors_impl.py", line 516, in exit
    c_api.TF_GetCode(self.status.status))
tensorflow.python.framework.errors_impl.InvalidArgumentError: Assign requires shapes of both tensors to match. lhs shape= [3,3,128,768] rhs shape= [5,5,128,768]
		[[Node: save/Assign_9 = Assign[T=DT_FLOAT, _class=["loc:@InceptionV4/AuxLogits/Conv2d_2a/weights"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/device:CPU:0"](InceptionV4/AuxLogits/Conv2d_2a/weights, save/RestoreV2:9)]]

During handling of the above exception, another exception occurred:
Traceback (most recent call last):
  File "compute_features.py", line 120, in <module>
    saver.restore(sess, checkpoint_file)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\training\saver.py", line 1775, in restore
    {self.saver_def.filename_tensor_name: save_path})
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\client\session.py", line 905, in run
    run_metadata_ptr)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\client\session.py", line 1140, in _run
    feed_dict_tensor, options, run_metadata)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\client\session.py", line 1321, in _do_run
    run_metadata)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\client\session.py", line 1340, in _do_call
    raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.InvalidArgumentError: Assign requires shapes of both tensors to match. lhs shape= [3,3,128,768] rhs shape= [5,5,128,768]
         [[Node: save/Assign_9 = Assign[T=DT_FLOAT, _class=["loc:@InceptionV4/AuxLogits/Conv2d_2a/weights"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/device:CPU:0"](InceptionV4/AuxLogits/Conv2d_2a/weights, save/RestoreV2:9)]]
Caused by op 'save/Assign_9', defined at:
  File "compute_features.py", line 119, in <module>
    saver = tf.train.Saver()
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\training\saver.py", line 1311, in init
    self.build()
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\training\saver.py", line 1320, in build
    self.build(self.filename, build_save=True, build_restore=True)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\training\saver.py", line 1357, in _build
    build_save=build_save, build_restore=build_restore)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\training\saver.py", line 809, in _build_internal
    restore_sequentially, reshape)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\training\saver.py", line 470, in _AddRestoreOps
    assign_ops.append(saveable.restore(saveable_tensors, shapes))
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\training\saver.py", line 162, in restore
    self.op.get_shape().is_fully_defined())
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\ops\state_ops.py", line 281, in assign
    validate_shape=validate_shape)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\ops\gen_state_ops.py", line 64, in assign
    use_locking=use_locking, name=name)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\framework\op_def_library.py", line 787, in _apply_op_helper
    op_def=op_def)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\framework\ops.py", line 3290, in create_opop_def=op_def)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\framework\ops.py", line 1654, in init
    self.traceback = self.graph._extract_stack()  # pylint: disable=protected-access
InvalidArgumentError (see above for traceback): Assign requires shapes of both tensors to match. lhs shape= [3,3,128,768] rhs shape= [5,5,128,768]
         [[Node: save/Assign_9 = Assign[T=DT_FLOAT, _class=["loc:@InceptionV4/AuxLogits/Conv2d_2a/weights"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/device:CPU:0"](InceptionV4/AuxLogits/Conv2d_2a/weights, save/RestoreV2:9)]]

[tensorflowbutler]

Thank you for your post. We noticed you have not filled out the following field in the issue template. Could you update them if they are relevant in your case, or leave them as N/A? Thanks.
Have I written custom code
OS Platform and Distribution
TensorFlow installed from
Bazel version
CUDA/cuDNN version
GPU model and memory
Exact command to reproduce

[zacharyclam]

I have added this information。

[reedwm]

This question is better asked on StackOverflow since it is not a bug or feature request. There is also a larger community that reads questions there.

If you think we've misinterpreted a bug, please comment again with a clear explanation, as well as all of the information requested in the issue template. Thanks!