ppplinday/tensorflow-vgg16-train-and-test

The purpose of this program is for studying. Using tensorflow trains the vgg16 and recognizes only two kinds of picture(cat and dog).

tensorflow-vgg16-train-and-test's Issues

about train

when i run the train.py, there are sth wrong with the codes,

training

Hello! I would like to ask if your training code can be used to train ILSVRC-2012 dataset？ Thank you！

about test

when I run test_vgg.py, I did not get the predict result, but some platform information such as "linux --python 3.6.8 ..."also it did not give a wrong or warning. I want to know why it happened, t hanks!

A Deep CNN based transfer learning method

CT image
Imagenet
vgg16
weights
svm
soft max
knn

How can this sample improve its accuracy?

First, thank you for sharing your codes.
And after i used your codes, i found that the accuracy is very low, what else can i do to improve it?

why I got the same loss and acc

I just change the pictures but got the loss and acc are same..

Something wrong after running test_vgg.py

Hi ,when i run test_vgg.py,Something wrong happend,
File "test_vgg.py", line 11, in recognize
output_graph_def.ParseFromString(f.read())
google.protobuf.message.DecodeError: Error parsing message

How to solve this problem ?

network question

I haven't find dropout function in your network。

question about 'softmax' in train_vgg.py

Hello there!

Your code is pretty clear and helpful to others, thank you for your effort!

Here, I raised a question about 'softmax', u write about this:

train_vgg.py:

#fc8
    with tf.name_scope('fc8') as scope:
        kernel = weight_variable([4096, 2])
        biases = bias_variable([2])
        output_fc8 = tf.nn.relu(fc(output_fc7, kernel, biases), name=scope)

    finaloutput = tf.nn.softmax(output_fc8, name="softmax")

    cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=finaloutput, labels=y))

Here, u used tf.nn.softmax_cross_entropy_with_logits() function, which I thought is inner implemented softmax to normalize data and then calcuate cross entropy, applied on finaloutput which has already been applied by 'softmax'. So is there more reasonable to use:

cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=output_fc8, labels=y))

?

That's to say, use output_fc8 result to calculate cross entropy.

IndexError: tuple index out of range

Hello, I test with your code for 10 categories images classification, but I get an error, do you know how to fix it?

I think in order to train a model for more than 2 categories, I need to modifiy some variables or parameters. Could you give me some hints? Thanks a lot.

Traceback (most recent call last):

  File "<ipython-input-6-96478bbedd1c>", line 271, in <module>
    main()

  File "<ipython-input-6-96478bbedd1c>", line 268, in main
    train_network(g, batch_size, num_epochs, pb_file_path)

  File "<ipython-input-6-96478bbedd1c>", line 254, in train_network
    total_cost_in_train_set))

IndexError: tuple index out of range