how to test the trained module? about cnn-text-classification-tf HOT 78 CLOSED

Yeah, I should probably write more extensive documentation and a word2vec example. I didn't expect this many people would use the code. I'll try to put something together on the weekend.

from cnn-text-classification-tf.

@JamesRedfield Thanks for finding this. It seems like most_common method does random sorting for words that occur the same number of times. So words may receive a different index during training and eval, resulting in wrong predictions.

I fixed this in 6727296

Let me know if it works! (you need to retrain)

from cnn-text-classification-tf.

Can someone share the syntax they used to test their own content (line 30 in eval.py)? I am curious about testing one sentence and also testing a batch of sentences.

from cnn-text-classification-tf.

Hello,
I tried running the eval.py giving the path to checkpoint folder, and getting the following error.
I am new at neural networks, could you please look into the error?

Traceback (most recent call last):
  File "eval.py", line 68, in <module>
    batch_predictions = sess.run(predictions, {input_x: x_test_batch, dropout_keep_prob: 1.0})
  File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 315, in run
    return self._run(None, fetches, feed_dict)
  File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 511, in _run
    feed_dict_string)
  File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 564, in _do_run
    target_list)
  File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 586, in _do_call
    e.code)
tensorflow.python.framework.errors.InvalidArgumentError: computed output size would be negative
     [[Node: conv-maxpool-5/pool = MaxPool[ksize=[1, 55, 1, 1], padding="VALID", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"](conv-maxpool-5/relu)]]
Caused by op u'conv-maxpool-5/pool', defined at:
  File "eval.py", line 50, in <module>
    saver = tf.train.import_meta_graph("{}.meta".format(checkpoint_file))
  File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/tensorflow/python/training/saver.py", line 1285, in import_meta_graph
    return _import_meta_graph_def(_read_meta_graph_file(meta_graph_or_file))
  File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/tensorflow/python/training/saver.py", line 1220, in _import_meta_graph_def
    importer.import_graph_def(meta_graph_def.graph_def, name="")
  File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/tensorflow/python/framework/importer.py", line 238, in import_graph_def
    compute_shapes=False, compute_device=False)
  File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 2040, in create_op
    original_op=self._default_original_op, op_def=op_def)
  File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 1087, in __init__
    self._traceback = _extract_stack()

from cnn-text-classification-tf.

PLEASE take me off this chain. Thanks.

On Monday, June 13, 2016, Denny Britz [email protected] wrote:

@rajmiglani https://github.com/rajmiglani See #10
#10 and
like zhang-jian said, make sure your vocabulary matches up.

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
#8 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/ALRpG44KMJRuCQXZRICRm7Iw1EKNoBNxks5qLfdsgaJpZM4HxCMG
.

Areeba Kamal
Mount Holyoke College '16

from cnn-text-classification-tf.

@JamesRedfield You can use the GPU, just not for the embeddings. Tensorflow is is missing support for the necessary operations. Maybe they fixed it in the latest version so you cold try to upgrade.

The rest should work fine on the GPU, are you sure it's set up correctly? Do the other Tensorflow examples run on it? Running TF on AWS GPU instances is a bit tricky, but a lot has been written about it online.

from cnn-text-classification-tf.

@JamesRedfield I'm not sure, according to tensorflow/tensorflow#305 it seems like they added support for it. So I guess it should work...

Your error Ran out of memory trying to allocate 1.62MiB. See logs for memory state doesn't really make sense to me. Is there no memory left on the GPU? It seems like other people have had similar issues: tensorflow/tensorflow#398

This seems like more of a Tensorflow issue, not an issue with this code. So maybe ask in the Tensorflow group and print out the detailed debug logs..

from cnn-text-classification-tf.

@JamesRedfield You are iterating through the training files multiple times anyway, just duplicating the data won't do anything (assuming you train long enough)

from cnn-text-classification-tf.

Yes, it's not language specific and should work as long as the language is not too different (like Chinese where you probably need to train on cahracters instead of words). Make sure that the vocabulary is big enough and that your test set doesn't use completely different vocabulary.

Btw, the convolutions are applied to the embedded vectors, not the one-hot vectors.

from cnn-text-classification-tf.

@aigujin Easiest way is probably to use http://scikit-learn.org/stable/modules/generated/sklearn.metrics.precision_recall_fscore_support.html with the predictions.

from cnn-text-classification-tf.

One more thing @dennybritz if i try to test it on another Q which has no word as in the trained vocablary , will it give me any errors while running or it will just reduce the accuracy?(right now i am just testing with the same words)

With the current code I believe it would throw an error. Here's how you usually deal with this:

1. Decide on a number of words in your vocabulary, e.g. 10k.
2. During data loading (in helpers.py), only consider the most common 10k words. Replace all other words with a special OOV (out of vocabulary) token.
3. During training you'll learn a representation for your 10k words and the OOV "word".
4. During testing do the same thing and replace all unknown words with the OOV token. So, that'll decrease your accuracy but won't throw an error.

If that's too much manual work, check out the VocabularyProcessor that comes with Tensorflow/skflow. It handles that for you. You can find an example here: https://github.com/tensorflow/tensorflow/blob/master/tensorflow/examples/skflow/text_classification.py

from cnn-text-classification-tf.

To be more precise, since the function build_vocab() learns the vocabulary mapping based on the input data, both training/evaluation/test data should be indexed using the same vocabulary mapping when replacing the OOV. Otherwise, the index number of the same word could be different between training/testing.

Also, you need to make sure the max sentence length is the actual max sentence length in all training/evaluation/test data sets.

from cnn-text-classification-tf.

Summaries are what you can visualize in Tensorboard. To load the trained model you need to load the checkpoint files. Check out the official documentation on that: https://www.tensorflow.org/versions/r0.7/how_tos/variables/index.html#variables-creation-initialization-saving-and-loading

from cnn-text-classification-tf.

@dennybritz
hi, there:
you use tf.train.Saver() to save all variables in train.py
but your W is defined in text_cnn.py
does the tensorflow save these variables
also the vocabulary, vocabulary_inv that is not defined by using tf.Variable()
does the tensorflow save these?
thanks in advance~

from cnn-text-classification-tf.

@zishell Check out the Tensorflow documentation for storing and saving variables: https://www.tensorflow.org/versions/r0.7/how_tos/variables/index.html#variables-creation-initialization-saving-and-loading

In short, all the model parameters (W, etc) are saved by Tensorflow because they are part of the implicit graph. However, the vocabulary isn't saved, you need to load it yourself. If it's not clear I recommend reading the Tensorflow documentation, it explains how variables and graphs are working.

from cnn-text-classification-tf.

@dennybritz thanks, I have understood of save and restore.
This is my script to classify one sentence:

if __name__ == '__main__':
    with tf.Session() as sess:

        # process the raw sentence 
        new_review = "simplistic , silly and tedious . "
        new_review = new_review.strip()
        new_review = data_helpers.clean_str(new_review)
        new_review = new_review.split(" ")

        sentences, dump= data_helpers.load_data_and_labels()
        sequence_length = max(len(x) for x in sentences)
        sentences_padded = data_helpers.pad_sentences(sentences)
        vocabulary, vocabulary_inv = data_helpers.build_vocab(sentences_padded)

        num_padding = sequence_length - len(new_review)
        new_sentence = new_review + ["<PAD/>"] * num_padding

        #convert x 
        x = np.array([vocabulary[word] for word in new_sentence])
        input_x = (x)


        sequence_length = x.shape[0]
        vocab_size = len(vocabulary)
        embedding_size = FLAGS.embedding_dim
        filter_sizes = map(int, FLAGS.filter_sizes.split(","))
        num_filters = FLAGS.num_filters
        l2_reg_lambda = FLAGS.l2_reg_lambda


        w_embedding = tf.Variable(tf.random_uniform([vocab_size, embedding_size], -1.0, 1.0), name="embedding/W")
        embedded_chars = tf.nn.embedding_lookup(w_embedding, input_x)
        embedded_chars_expanded = tf.expand_dims(embedded_chars, -1)
        # Restore variables from disk.
        saver = tf.train.Saver()
        saver.restore(sess, "runs/1459166181/checkpoints/model-20000")
        print("Model restored.")

        # Create a convolution + maxpool layer for each filter size
        pooled_outputs = []
        for i, filter_size in enumerate(filter_sizes):
            with tf.name_scope("conv-maxpool-%s" % filter_size):
                # Convolution Layer
                filter_shape = [filter_size, embedding_size, 1, num_filters]
                W = tf.Variable(tf.truncated_normal(filter_shape, stddev=0.1), name="W")
                b = tf.Variable(tf.constant(0.1, shape=[num_filters]), name="b")
                # Restore variables from disk.
                saver = tf.train.Saver()
                saver.restore(sess, "runs/1459166181/checkpoints/model-20000")
                print("Model restored.")


                conv = tf.nn.conv2d(
                    embedded_chars_expanded,
                    W,
                    strides=[1, 1, 1, 1],
                    padding="VALID",
                    name="conv")
                # Apply nonlinearity
                h = tf.nn.relu(tf.nn.bias_add(conv, b), name="relu")
                # Maxpooling over the outputs
                pooled = tf.nn.max_pool(
                    h,
                    ksize=[1, sequence_length - filter_size + 1, 1, 1],
                    strides=[1, 1, 1, 1],
                    padding='VALID',
                    name="pool")
                pooled_outputs.append(pooled)

        # Combine all the pooled features
        num_filters_total = num_filters * len(filter_sizes)
        h_pool = tf.concat(3, pooled_outputs)
        h_pool_flat = tf.reshape(h_pool, [-1, num_filters_total])

        # Add dropout
        h_drop = tf.nn.dropout(h_pool_flat, 0.5)

        # Final (unnormalized) scores and predictions
        W = tf.Variable(tf.truncated_normal([num_filters_total, num_classes], stddev=0.1), name="output/W")
        b = tf.Variable(tf.constant(0.1, shape=[num_classes]), name="output/b")
        # Restore variables from disk.
        saver = tf.train.Saver()
        saver.restore(sess, "runs/1459166181/checkpoints/model-20000")
        print("Model restored.")


        scores = tf.nn.xw_plus_b(h_drop, W, b, name="scores")
        predictions = tf.argmax(scores, 1, name="predictions")
        print predictions

I have two questions:

1. How to convert x to input_x, when I debug the train.py, found that x_batch is tuple, so I convert the x to (x), is that right

w_embedding = tf.Variable(tf.random_uniform([vocab_size, embedding_size], -1.0, 1.0), name="embedding/W")
embedded_chars = tf.nn.embedding_lookup(w_embedding, input_x)
embedded_chars_expanded = tf.expand_dims(embedded_chars, -1)

the dimension of embedded_chars is [56,128] not like yours [?,56,128]
and the dimension of embedded_chars_expanded is [56,128,1] not like yours [?,56,128,1]

How can I convert it, thanks

from cnn-text-classification-tf.

@dennybritz I've read the Tensorflow documentation, but still couldn't quite get it right.

Since many people have asked similar question in the tutorial's comments section.
Is it possible for you to update the sample code with an eval example?

from cnn-text-classification-tf.

Yeah, maybe that's a good idea. I'll try to get it done this weekend.

from cnn-text-classification-tf.

@dennybritz It will be very helpful if you share the work you doing now as soon as possible..thanks in advance

@zishell I tried your code for testing that particular sentence, but i got an error like

NameError:

name 'FLAGS' is not defined

How can I resolve this?

from cnn-text-classification-tf.

@AAMIBhavya I did not paste the head of my script, you should add the flags by yourself copy from the train.py

from cnn-text-classification-tf.

Okay, I added a sample eval.py. Also added a note to the README. Let me know if this works.

EDIT: Also, make sure you pull the latest code changes, I added an additional flag to the data helpers that's needed for the eval.

from cnn-text-classification-tf.

@dennybritz Thanks, when I run eval.py, I got output like this
`BATCH_SIZE=64
CHECKPOINT_DIR=/home/product/Downloads/cnn-text-classification-tf-master/runs/1459745478/checkpoints/
LOG_DEVICE_PLACEMENT=False

Loading data...
Vocabulary size: 18765
Test set size 10662

Evaluating...

Total number of test examples: 10662
Accuracy: 0
product@product:~/Downloads/cnn-text-classification-tf-master$ python eval.py

Parameters:
ALLOW_SOFT_PLACEMENT=True
BATCH_SIZE=64
CHECKPOINT_DIR=/home/product/Downloads/cnn-text-classification-tf-master/runs/1459745478/checkpoints/
LOG_DEVICE_PLACEMENT=False

Loading data...
[[ 1 571 7 ..., 0 0 0]
[ 1 3805 2181 ..., 0 0 0]
[ 718 13 44 ..., 0 0 0]
...,
[ 12 9 1474 ..., 0 0 0]
[ 1 166 439 ..., 0 0 0]
[3308 7 63 ..., 0 0 0]]
[1 1 1 ..., 0 0 0]
Vocabulary size: 18765
Test set size 10662

Evaluating...

Total number of test examples: 10662
Accuracy: 0`

Why I got accuracy as 0 (I just printed the values of x_test and y_test also). And one more thing, how can I test a sentence which is not in training sample, since it doesn't have any value for y_test know?

from cnn-text-classification-tf.

@zishell Thanks. I corrected that part. But I got error after restoring the trained model as

**Model restored.
Model restored.
Traceback (most recent call last):
File "test.py", line 90, in
name="conv")
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/gen_nn_ops.py", line 211, in conv2d
use_cudnn_on_gpu=use_cudnn_on_gpu, name=name)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/op_def_library.py", line 655, in apply_op
op_def=op_def)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 2042, in create_op
set_shapes_for_outputs(ret)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 1528, in set_shapes_for_outputs
shapes = shape_func(op)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/common_shapes.py", line 176, in conv2d_shape
input_shape = op.inputs[0].get_shape().with_rank(4)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/tensor_shape.py", line 616, in with_rank
raise ValueError("Shape %s must have rank %d" % (self, rank))
ValueError: Shape (56, 128, 1) must have rank 4

What this value error corresponds to? How can I resolve this?

from cnn-text-classification-tf.

@AAMIBhavya You may need to pull the updated code and retrain the model. The eval script I wrote only works with Tensorflow 0.7, so make sure you have the latest version.

For testing a single sentence, it should be easy to modify the eval script for that. Just change the data loading section to load your own sentence and then remove the accuracy calculation and print out the prediction instead.

from cnn-text-classification-tf.

@dennybritz Thanks. Actually the version of Tensorflow I am using is 0.7.1

from cnn-text-classification-tf.

Hi guys! Anyone has understood how can use eval.py to test a single sentence ( present in training sample ). I just change the data loading section to load one sentence. but I feel confused when I try to print out the prediction.

• after Load my sentence

• after add pad to my sentence to the same length of the longest sentence in sample data

• after mapping sentence to vectors based on a vocabulary. ( data vocabulary from sample data )

• after generate batches for one epoch

  I 'm not able to understand how print out prediction. This is my output

Loading data...
Vocabulary size: 18762
Test set size 10662

Evaluating...

[[1207 2 304 4 660 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0]]

It corresponds to "simplistic , silly and tedious ."
Now i don't know how use y_test fro prediction. Please help me

from cnn-text-classification-tf.

@AAMIBhavya It seems strange that you would get an accuracy of 0. I just tested this locally and it's working for me if I train and then point to the checkpoint dir. I'm using the latest version of Tensorflow and Python 3.

@kinarashah I haven't seen this before, but could you try train + eval with Python 3 and the latest version of Tensorflow?

@JamesRedfiled You don't need the y_test for predictions. These are the labels and they are only used to calculate the accuracy. To get the predictions you can just print out the all_predictions variable.

from cnn-text-classification-tf.

@dennybritz Thanks a lot! I just tried to print 'all predictions' with always the same input sentence..( exactly sentence pick up from sample data rt-polarity.neg ). The response is never equal..

1 ) Sometimes return 1 or 0. It depends from different representations of the same input. Is it right ?

2 ) Why the representation of the same sentence changes every time? The "build_vocab(sentences)" function changes at every execution. Why?

3. Is it depends by mapping of words with same most_common count value?

For example - the first word 'simplistic' is mapped like: 1192, then 1237 or then 1191. How is possible this little difference generates a totally difference response at the end ( positive or negative sentence)?

There are two output response from the identical input sentence ( "simplistic , silly and tedious ." )

$ python3.5 eval.py

Loading data...
Vocabulary size: 18762
Test set size 10662

sentence - Mapped sentence to vector based on a vocabulary.
[[1192 2 303 4 666 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0]]

Evaluating...
[ 1.]

$ python3.5 eval.py

Loading data...
Vocabulary size: 18762
Test set size 10662

sentence - Mapped sentence to vector based on a vocabulary.
[[1237 2 308 4 661 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0]]

Evaluating...
[ 0.]

Thank you for your time Denny!

from cnn-text-classification-tf.

@JamesRedfield Thanks, I'll look into it today and get back to you. The predictions should be equal, so something may be wrong there.

from cnn-text-classification-tf.

@dennybritz Thanks for your quickly answer! I used your code after retrain data. The output have more sense now but i tested your algorithm with all negative sentences ( from sample data ) in input:

I am aspected almost all same prediction [0.] instead the response is ( more Zero and some Ones ). - Is it depend from size of dataset? Is it right?

• Can I have all negative predictions when I using exactly negative sentences from sample data?
• Do you think the positive predictions are too much?

Output:

Loading data...
Vocabulary size: 18765
Input all negative Sentences : 5331

Evaluating...
Positives: 176
Negatives: 5155

We have 3,3% of not true response.

I hope my explain is clear.
Thanks!

from cnn-text-classification-tf.

@dennybritz When we use new test data, there may be new terms which is not in the vocabulary. How it can be handled? Can that new terms also include in that vocabulary?

from cnn-text-classification-tf.

Hi @AAMIBhavya. In my case, the solution is to ignore the new terms because is not mapped. If you want to include the new terms, you 'll need to retrain the data set with new sentences with new terms.

from cnn-text-classification-tf.

@JamesRedfield Thanks..

from cnn-text-classification-tf.

@AAMIBhavya Yes, the most common way is to limit the vocabulary to a certain number of most common words (e.g. 10k or 20k) and replace all other words with a special UNK word/token, both during training and test. All words not in the vocab are then treated the same.

A more sophisticated approach is to use character-level models to "generalize" to new words, e.g. http://arxiv.org/abs/1602.00367

from cnn-text-classification-tf.

Hi. I also tried the eval script but I have the same problem of @kinarashah

`W tensorflow/core/kernels/pooling_ops_common.cc:64] Invalid argument: computed output size would be negative
W tensorflow/core/common_runtime/executor.cc:1102] 0x2bcda30 Compute status: Invalid argument: computed output size would be negative
[[Node: conv-maxpool-5/pool = MaxPoolksize=[1, 32, 1, 1], padding="VALID", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"]]
W tensorflow/core/kernels/pooling_ops_common.cc:64] Invalid argument: computed output size would be negative
W tensorflow/core/common_runtime/executor.cc:1102] 0x2bcda30 Compute status: Invalid argument: computed output size would be negative
[[Node: conv-maxpool-4/pool = MaxPoolksize=[1, 33, 1, 1], padding="VALID", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"]]
W tensorflow/core/kernels/pooling_ops_common.cc:64] Invalid argument: computed output size would be negative
W tensorflow/core/common_runtime/executor.cc:1102] 0x2bcda30 Compute status: Invalid argument: computed output size would be negative
[[Node: conv-maxpool-3/pool = MaxPoolksize=[1, 34, 1, 1], padding="VALID", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"]]
Traceback (most recent call last):
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/client/session.py", line 571, in _do_call
return fn(*args)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/client/session.py", line 555, in _run_fn
return tf_session.TF_Run(session, feed_dict, fetch_list, target_list)
tensorflow.python.pywrap_tensorflow.StatusNotOK: Invalid argument: computed output size would be negative
[[Node: conv-maxpool-5/pool = MaxPoolksize=[1, 32, 1, 1], padding="VALID", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"]]

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "eval.py", line 69, in
batch_predictions = sess.run(predictions, {input_x: x_test_batch, dropout_keep_prob: 1.0})
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/client/session.py", line 315, in run
return self._run(None, fetches, feed_dict)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/client/session.py", line 511, in _run
feed_dict_string)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/client/session.py", line 564, in _do_run
target_list)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/client/session.py", line 586, in _do_call
e.code)
tensorflow.python.framework.errors.InvalidArgumentError: computed output size would be negative
[[Node: conv-maxpool-5/pool = MaxPoolksize=[1, 32, 1, 1], padding="VALID", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"]]
Caused by op 'conv-maxpool-5/pool', defined at:
File "eval.py", line 51, in
saver = tf.train.import_meta_graph("{}.meta".format(checkpoint_file))
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/training/saver.py", line 1285, in import_meta_graph
return _import_meta_graph_def(_read_meta_graph_file(meta_graph_or_file))
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/training/saver.py", line 1220, in _import_meta_graph_def
importer.import_graph_def(meta_graph_def.graph_def, name="")
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/framework/importer.py", line 238, in import_graph_def
compute_shapes=False, compute_device=False)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/framework/ops.py", line 2040, in create_op
original_op=self._default_original_op, op_def=op_def)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/framework/ops.py", line 1087, in init
self._traceback = _extract_stack()`

I tried both with Python 2.7 and Python 3.4 and I have the same problem.
How I can fix?

from cnn-text-classification-tf.

@fmaglia I'm not sure what could be wrong there. I haven't seen this error before.

Are you using the latest version of Tensorflow? What are you running on (Linux/Windows/Mac)? Are you using your own data are only the code here?

from cnn-text-classification-tf.

I'm using the latest version of Tensorflow.
I'm running on Linux Ubuntu 14.04 LTS.
I'm using my own data (tweet of SemEval competition).
Thanks for the quick reply.

dataset.zip

from cnn-text-classification-tf.

Does the code work for you with the default data here? Are you sure you have the same vocabulary for both training and test data, and that the test data doesn't contain any out of vocabulary words?

from cnn-text-classification-tf.

I didn't tried the code with the default data.
I have different vocabulary for training and test data because the dataset are different. One is for the training of the classifier, the second is for the testing of the neural network.
Why the vocabulary should be the same?

from cnn-text-classification-tf.

The vocabulary must be the same or you can't index the embeddings. That's standard for almost all models (other than those that explicitly generalize to new terms, like character models). It has been discussed above, see #8 (comment)

Limit the vocabulary to the N most common terms and use them for your train/test.

from cnn-text-classification-tf.

@AAMIBhavya I was having issues with the eval script saying that accuracy was 0 as well, but I went in and changed correct_predictions and len(y_test) to floats instead of ints and that solved the problem for me.

from cnn-text-classification-tf.

@NickLevitt That makes sense, thanks for figuring that out. I changed it to float in the code now. Strange that it worked for me without though.

from cnn-text-classification-tf.

@dennybritz Thanks for your quick replies..I corrected all the errors and I also tested with my own test corpus..it gave good results. Good work..
@NickLevitt Thanks..I was having that accuracy 0 issue, but I discover that it was because of python 2. When I run that code using python3 I got the correct value for accuracy. That *_accuracy 0 *_problem comes only when you use python 2 (since there is a need for type casting). So try with python 3..

from cnn-text-classification-tf.

@dennybritz I'm also facing the same issue as @fmaglia . What I did to evaluate, was to keep just the first sentence in both the pos and neg files and run the eval.py for this data.

The code works perfectly fine for the complete dataset and gives accuracy of 0.973551 as mentioned by you. Please let me know if you have any pointers for me. Thanks.

from cnn-text-classification-tf.

@nishantvishwamitra You must have the same vocabulary during training and test. In eval, you need to build the full vocabulary from the training data, and then map your test sentences in that vocabulary.

from cnn-text-classification-tf.

I try to solve my problem using the vocabulary of the training test also for the test set and removing words of test set that aren't in the training set but when I launch the eval script :
`W tensorflow/core/kernels/pooling_ops_common.cc:64] Invalid argument: computed output size would be negative
W tensorflow/core/common_runtime/executor.cc:1102] 0x18a1fb0 Compute status: Invalid argument: computed output size would be negative
[[Node: conv-maxpool-3/pool = MaxPoolksize=[1, 34, 1, 1], padding="VALID", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"]]
W tensorflow/core/kernels/pooling_ops_common.cc:64] Invalid argument: computed output size would be negative
W tensorflow/core/common_runtime/executor.cc:1102] 0x18a1fb0 Compute status: Invalid argument: computed output size would be negative
[[Node: conv-maxpool-5/pool = MaxPoolksize=[1, 32, 1, 1], padding="VALID", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"]]
W tensorflow/core/kernels/pooling_ops_common.cc:64] Invalid argument: computed output size would be negative
W tensorflow/core/common_runtime/executor.cc:1102] 0x18a1fb0 Compute status: Invalid argument: computed output size would be negative
[[Node: conv-maxpool-4/pool = MaxPoolksize=[1, 33, 1, 1], padding="VALID", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"]]
Traceback (most recent call last):
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/client/session.py", line 571, in _do_call
return fn(*args)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/client/session.py", line 555, in _run_fn
return tf_session.TF_Run(session, feed_dict, fetch_list, target_list)
tensorflow.python.pywrap_tensorflow.StatusNotOK: Invalid argument: computed output size would be negative
[[Node: conv-maxpool-3/pool = MaxPoolksize=[1, 34, 1, 1], padding="VALID", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"]]

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "eval.py", line 70, in
batch_predictions = sess.run(predictions, {input_x: x_test_batch, dropout_keep_prob: 1.0})
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/client/session.py", line 315, in run
return self._run(None, fetches, feed_dict)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/client/session.py", line 511, in _run
feed_dict_string)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/client/session.py", line 564, in _do_run
target_list)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/client/session.py", line 586, in _do_call
e.code)
tensorflow.python.framework.errors.InvalidArgumentError: computed output size would be negative
[[Node: conv-maxpool-3/pool = MaxPoolksize=[1, 34, 1, 1], padding="VALID", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"]]
Caused by op 'conv-maxpool-3/pool', defined at:
File "eval.py", line 51, in
saver = tf.train.import_meta_graph("{}.meta".format(checkpoint_file))
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/training/saver.py", line 1285, in import_meta_graph
return _import_meta_graph_def(_read_meta_graph_file(meta_graph_or_file))
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/training/saver.py", line 1220, in _import_meta_graph_def
importer.import_graph_def(meta_graph_def.graph_def, name="")
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/framework/importer.py", line 238, in import_graph_def
compute_shapes=False, compute_device=False)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/framework/ops.py", line 2040, in create_op
original_op=self._default_original_op, op_def=op_def)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/framework/ops.py", line 1087, in init
self._traceback = _extract_stack()
`
Have you any idea about what causing error?

from cnn-text-classification-tf.

Hi @dennybritz ! The embedding implementation doesn’t currently have GPU support and throws an error if placed on the GPU.

Is it means that I can't use GPU anyway? I tried your algorithm on AMI and I'm' trying to figure that out.. but the result is:

W tensorflow/core/common_runtime/gpu/gpu_bfc_allocator.cc:211] Ran out of memory trying to allocate 1.62MiB. See logs for memory state

W ./tensorflow/core/kernels/assign_op.h:77] Resource exhausted: OOM when allocating tensor with shape[4,128,1,128]
W tensorflow/core/common_runtime/executor.cc:1102] 0x274a8c0 Compute status: Resource exhausted: OOM when allocating tensor with shape[4,128,1,128]

thanks

from cnn-text-classification-tf.

I believe all is correct.
I had some problem to use tensorflow 0.8 with CUDA 7.5 and cudnn 5.0. In my case, It was very tricky run a simple example like this:
import tensorflow as tf
a = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[2, 3], name='a')
b = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[3, 2], name='b')
c = tf.matmul(a, b)
sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
print sess.run(c)

to use GPU.

At the end, after I spent 3-4 days to read moooore forum I realized one configuration, tensor flow 0.7 - CUDA 7.0 - CUDNN 4.0. I run on python3.4. This is my case..

Now I run your code after reboot the server but I see something not correct. Your experience, in this case, is very precious.

This is the process
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
ubuntu 2177 143 2.4 39720844 383504 pts/0 Sl+ 14:14 8:26 python3.4 train.py

Nvidia information
NVIDIA-SMI 346.46 Driver Version: 346.46 |
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 2177 C python3.4 3819MiB |

but the training time is exactly the same. I'm using GPU device
with tf.device('/gpu:0'), tf.name_scope("embedding"):

Do you know where I wrong? Any ideas?

I forgot it..Thanks for your support

from cnn-text-classification-tf.

As I mentioned above, you can't use the GPU for the embeddings with tf.device('/gpu:0'), tf.name_scope("embedding"): - you need to use the CPU (at least that was case with TF a while ago).

You can run the rest of the network on the GPU.

from cnn-text-classification-tf.

Oh sorry! I understand ! I believed the 0.7 tensorflow is relative new to resolve this trouble but when you said Maybe they fixed in the "latest version" , you referred to 0.8. Thanks a lot!

from cnn-text-classification-tf.

I'll look around before changing my "safe" setting. :) Thank for you incredible support @dennybritz

from cnn-text-classification-tf.

Hi @dennybritz.
If I have, for example, positive rt-polarity data file with 5000 lines and negative rt-polarity file with 2000 lines... And if I duplicate the negative lines to reach 5000 lines ( like positive input ). How change the results or the quality of algorithm. Is it still reliable?

from cnn-text-classification-tf.

Hello @dennybritz. I have another question and I hope not to be tedious. :) The algorithm does not care about a specific language? I mean, you do not used pre-trained word2vec vectors for word embeddings. Instead, you learn embeddings from scratch. You apply convolutions directly to one-hot vectors so the algorithm is not specific natural language oriented. Is it right?
I trained more data in a different language( 10.000 lines - 5000 pos - 5000 neg). At the end the response is not good. I mean, If I submit a new set of negative sentences, similar to the negative trained sentences, the response is statistically in the middle so half is classified like negative and half like positive. Do you known why? What do you think?

from cnn-text-classification-tf.

Hi @dennybritz

The vocabolary size is: 13K
The number of sentences is: 9000 ( 4500 are positive - 1200 are negative [4500 with duplication] )

Test set:
new negative sentences are 60
words: 1100 of which 290 are no mapped
So the average of words not mapped, in new sentences, is 30%.

But in some cases, the number of words not mapped in a new NEGATIVE sentence, is zero, so the response is the same POSITIVE. Is it makes sense for you?

The result of elaboration is half sentences classified like positive and half negative.I tried with an different test set size ( like 300 new negative sentences) but the result is exactly the same.

What do you think? Is the vocabulary big enough?

ps. when I discover a new word not mapped, during eval.py elaboration I put a PAD character like a value
thanks

from cnn-text-classification-tf.

@JamesRedfield This seems reasonable to me, not sure what could be wrong there.

from cnn-text-classification-tf.

@JamesRedfield / @AAMIBhavya /@dennybritz : do you have sample code for loading your own single sentence to detect sentiment? I'm running this for a class and it would be a huge help to see it. Thanks.

from cnn-text-classification-tf.

Hi, thanks for the code. Just one question: how to modify (if possible at all) eval.py to get a precision/recall table?

from cnn-text-classification-tf.

@dennybritz Hi, thanks for the suggestion; it worked. Now, I would like to clarify a couple of issues:

1. If I have out-of-sample sentences to predict classes, I ignore the variable y_test, right? (I don't know the classes and want a model to classify)
2. In the same set up (out-of-sample), do I need to build a new vocabulary or I can use the one from the model?

from cnn-text-classification-tf.

@aigujin

1. Yes
2. You must to use the same vocabulary you used during training because of the learned word mappings. You don't build a new one. If your examples are very different types of sentences than those used in training it may be better to re-train.

from cnn-text-classification-tf.

Hi @dennybritz I have a question for you!
I modify the code for training and evaluation the same data in every simulation because I would like to analyze the changing of the accuracy in the classification of the dev set. I also modified the eval.py, but the result of the eval script is different to the result of the dev accuracy in the train script. In "eval.py" the result about accuracy is better than "train.py".
I didn't modified the tensorflow code, but the only the operations in the data_helpers.py and the inizialization of the data in train.py and the eval.py.
Have you any ideas about this situation?

from cnn-text-classification-tf.

@fmaglia The eval.py file loads the training data by default. It's expected to get better accuracy on the training data than on dev/test data (obviously, since you're training on it). If you have a test set you should load that in eval.py.

from cnn-text-classification-tf.

I splitted the initial dataset in training set (67% of the dataset) and in dev/test data (33% of the dataset).
In train.py I used the training set for the training and the test set for the classification, instead in eval.py I used only the test set for the evaluation. The data set used are the same but not the results obtained on the evaluation/classification of the dev/test set!

from cnn-text-classification-tf.

That's very strange, it should be the same if you're using the same test set.

from cnn-text-classification-tf.

I found now the error. Thanks so much!

from cnn-text-classification-tf.

@fmaglia Hi, how did you solve the error in this comment. It seems, I am getting the same.
Thanks

from cnn-text-classification-tf.

@aigujin In that case I was trying to evaluate a test set that are composed by the words, that aren't all in the vocabulary.
I solved it by changing the data_helpers.py.
I inserted the test data in the training set and I split the dataset for training and a part for testing.
But I encountered a problem: the data_helpers script inserted positive sentences and negative sentences. The latest sentences (used for test set) are all negative!

I solved with this code (in data_helpers.py):
` # Generate labels
positive_labels = [[0, 1] for _ in positive_examples]
negative_labels = [[1, 0] for _ in negative_examples]
labels = []
totLabels = len(positive_examples) + len(negative_examples)
for j in range(totLabels):
if j%2==0:
labels.append([0,1])
else:
labels.append([1,0])

#ordering the dataset
x_text = []
totale = len(positive_examples) + len(negative_examples)
for i in range(totale):
    if i%2==0:
        x_text.append(positive_examples.pop(0))
    else:
        x_text.append(negative_examples.pop(0))`

P.S.
my dataset is balanced.

from cnn-text-classification-tf.

Hey,
I want to classify Q & A with 45 classes .Can i use this model ? I have already pre-processed the data into one-hot encoding (45 size vector) as was done in this post.What else do i have to change in the code?
I trained without any errors.My vocab size is only 550 words.
Also how to evaluate a single Q's class?I have also made the changes for the batch-size and epoch parameters as my data-set is pretty small consisting of only 356 Q with 45 classes.
I tried to run the eval.py changing the data file(in data_helpers.py) to a file with only one Q print(all_predictions) but i am getting certain errors similar to @fmaglia .Can someone be more specific how to proceed?
Also what train/dev should i keep.at present its 346/10( i know its a small data-set but want to check if it works)
Kindly reply as soon as possible.@dennybritz @JamesRedfield

two errors:

InvalidArgumentError Traceback (most recent call last)
in ()
52
53 for x_test_batch in batches:
---> 54 batch_predictions = sess.run(predictions, {input_x: x_test_batch, dropout_keep_prob: 1.0})
55 all_predictions = np.concatenate([all_predictions, batch_predictions])
56

/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc in run(self, fetches, feed_dict)
313 Tensor that doesn't exist.
314 """
--> 315 return self._run(None, fetches, feed_dict)
316
317 def partial_run(self, handle, fetches, feed_dict=None):

/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc in _run(self, handle, fetches, feed_dict)
509 # Run request and get response.
510 results = self._do_run(handle, target_list, unique_fetches,
--> 511 feed_dict_string)
512
513 # User may have fetched the same tensor multiple times, but we

/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc in _do_run(self, handle, target_list, fetch_list, feed_dict)
562 if handle is None:
563 return self._do_call(_run_fn, self._session, feed_dict, fetch_list,
--> 564 target_list)
565 else:
566 return self._do_call(_prun_fn, self._session, handle, feed_dict,

/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc in _do_call(self, fn, *args)
584 # pylint: disable=protected-access
585 raise errors._make_specific_exception(node_def, op, error_message,
--> 586 e.code)
587 # pylint: enable=protected-access
588 six.reraise(e_type, e_value, e_traceback)

and second same as @fmaglia

from cnn-text-classification-tf.

want to classify Q & A with 45 classes .Can i use this model ?

Yes, it should be the same.

I can't tell much based on the error you posted. It may help to post properly formatted code/errors and a diff with your changes. Change the num_classes argument when instantiating the TextCNN during training (see train.py) and make sure your data preprocessing is right. Make sure training works first, not eval.

from cnn-text-classification-tf.

Hey,
Training is working fine but the error occurs in eval.py.
I am posting the data_helpers.py file(changed only the function that processes data) along with my data-set(part of it) here, kindly help me resolve the errors as soon as possible.The changes done on my part are just to change the file in data_helpers.py to a file with just one Q. Vocabulary size being returned while eval.py is just the no of word in the new test file(7 in my case but these words were present while training the model with my own dataset, vocab size being 508 ).
Am i doing something wrong here? i have not done any change in eval.py but just changed the data_helpers.py to load my own data as mentioned.
i don't get denny's comment :
You must have the same vocabulary during training and test. In eval, you need to build the full vocabulary from the training data, and then map your test sentences in that vocabulary.

@fmaglia @dennybritz @JamesRedfield

errors:

---------------------------------------------------------------------------
InvalidArgumentError                      Traceback (most recent call last)
<ipython-input-25-e4c837cef39c> in <module>()
     56         for x_test_batch in batches:
     57             print()
---> 58             batch_predictions = sess.run(predictions, {input_x: x_test_batch, dropout_keep_prob: 1.0})
     59             all_predictions = np.concatenate([all_predictions, batch_predictions])
     60 

/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc in run(self, fetches, feed_dict)
    313         `Tensor` that doesn't exist.
    314     """
--> 315     return self._run(None, fetches, feed_dict)
    316 
    317   def partial_run(self, handle, fetches, feed_dict=None):

/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc in _run(self, handle, fetches, feed_dict)
    509     # Run request and get response.
    510     results = self._do_run(handle, target_list, unique_fetches,
--> 511                            feed_dict_string)
    512 
    513     # User may have fetched the same tensor multiple times, but we

/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc in _do_run(self, handle, target_list, fetch_list, feed_dict)
    562     if handle is None:
    563       return self._do_call(_run_fn, self._session, feed_dict, fetch_list,
--> 564                            target_list)
    565     else:
    566       return self._do_call(_prun_fn, self._session, handle, feed_dict,

/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.pyc in _do_call(self, fn, *args)
    584         # pylint: disable=protected-access
    585         raise errors._make_specific_exception(node_def, op, error_message,
--> 586                                               e.code)
    587         # pylint: enable=protected-access
    588       six.reraise(e_type, e_value, e_traceback)

InvalidArgumentError: computed output size would be negative
     [[Node: conv-maxpool-3/pool = MaxPool[ksize=[1, 21, 1, 1], padding="VALID", strides=[1, 1, 1, 1], _device="/job:localhost/replica:0/task:0/cpu:0"](conv-maxpool-3/relu)]]
Caused by op u'conv-maxpool-3/pool', defined at:
  File "/usr/lib/python2.7/runpy.py", line 162, in _run_module_as_main
    "__main__", fname, loader, pkg_name)
  File "/usr/lib/python2.7/runpy.py", line 72, in _run_code
    exec code in run_globals
  File "/usr/local/lib/python2.7/dist-packages/ipykernel/__main__.py", line 3, in <module>
    app.launch_new_instance()
  File "/usr/local/lib/python2.7/dist-packages/traitlets/config/application.py", line 589, in launch_instance
    app.start()
  File "/usr/local/lib/python2.7/dist-packages/ipykernel/kernelapp.py", line 405, in start
    ioloop.IOLoop.instance().start()
  File "/usr/local/lib/python2.7/dist-packages/zmq/eventloop/ioloop.py", line 162, in start
    super(ZMQIOLoop, self).start()
  File "/usr/local/lib/python2.7/dist-packages/tornado/ioloop.py", line 883, in start
    handler_func(fd_obj, events)
  File "/usr/local/lib/python2.7/dist-packages/tornado/stack_context.py", line 275, in null_wrapper
    return fn(*args, **kwargs)
  File "/usr/local/lib/python2.7/dist-packages/zmq/eventloop/zmqstream.py", line 440, in _handle_events
    self._handle_recv()
  File "/usr/local/lib/python2.7/dist-packages/zmq/eventloop/zmqstream.py", line 472, in _handle_recv
    self._run_callback(callback, msg)
  File "/usr/local/lib/python2.7/dist-packages/zmq/eventloop/zmqstream.py", line 414, in _run_callback
    callback(*args, **kwargs)
  File "/usr/local/lib/python2.7/dist-packages/tornado/stack_context.py", line 275, in null_wrapper
    return fn(*args, **kwargs)
  File "/usr/local/lib/python2.7/dist-packages/ipykernel/kernelbase.py", line 260, in dispatcher
    return self.dispatch_shell(stream, msg)
  File "/usr/local/lib/python2.7/dist-packages/ipykernel/kernelbase.py", line 212, in dispatch_shell
    handler(stream, idents, msg)
  File "/usr/local/lib/python2.7/dist-packages/ipykernel/kernelbase.py", line 370, in execute_request
    user_expressions, allow_stdin)
  File "/usr/local/lib/python2.7/dist-packages/ipykernel/ipkernel.py", line 175, in do_execute
    shell.run_cell(code, store_history=store_history, silent=silent)
  File "/usr/local/lib/python2.7/dist-packages/IPython/core/interactiveshell.py", line 2723, in run_cell
    interactivity=interactivity, compiler=compiler, result=result)
  File "/usr/local/lib/python2.7/dist-packages/IPython/core/interactiveshell.py", line 2825, in run_ast_nodes
    if self.run_code(code, result):
  File "/usr/local/lib/python2.7/dist-packages/IPython/core/interactiveshell.py", line 2885, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-25-e4c837cef39c>", line 37, in <module>
    saver = tf.train.import_meta_graph("{}.meta".format(checkpoint_file))
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/saver.py", line 1285, in import_meta_graph
    return _import_meta_graph_def(_read_meta_graph_file(meta_graph_or_file))
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/saver.py", line 1220, in _import_meta_graph_def
    importer.import_graph_def(meta_graph_def.graph_def, name="")
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/importer.py", line 238, in import_graph_def
    compute_shapes=False, compute_device=False)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 2040, in create_op
    original_op=self._default_original_op, op_def=op_def)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 1087, in __init__
    self._traceback = _extract_stack()

One more thing @dennybritz if i try to test it on another Q which has no word as in the trained vocablary , will it give me any errors while running or it will just reduce the accuracy?(right now i am just testing with the same words)

data_helpers.py function

def load_data_and_labels():
    """
    Loads  data from files, splits the data into words and generates labels.
    Returns split sentences and labels.i am able to get one-hot vectors with 45 classes as [1,0,0,0,0.....,0]
    """




    label=np.array([np.zeros(45)])
    lab=np.array([np.zeros(45)])

    #load data
    class_no=0


    x_texti=list(open("./data/file.txt","r").readlines())
    x_texti = [s.strip() for s in x_texti]
    length = len(x_texti)

    #generate labels
    lab[0][class_no]=1
    label=lab
    count=1
   # print(label)
    for i in range(1,length):

        if(x_texti[i]!="----"):
            label=np.concatenate((label,lab),0)

            count=count+1

            #print(i)
            #print(label)


        else:
            lab[0][class_no]=0
            class_no=class_no+1
            lab[0][class_no]=1
            continue

    #generate labels
    #label=[lab for _ in x_texti]
    #y=label
    #x_text=np.array([[x_texti[0]]])
    #split by words
    x_text=[]
    for sent in x_texti:
        if(sent!="----"):
            x_text=x_text+[clean_str(sent)]
    #split by words
    #x_text = [clean_str(sent) for sent in x_texti ]
    x_text = [s.split(" ") for s in x_text]

    return [x_text,label,count]

Data-set:

Does the Natural Language Classifier train using all of the data, or does it partition it in some way?
What kind of preprocessing does the classifier perform on its input?
Does the classifier train using all of the data, or does it partition it in some way?
does the classifier train on all the data or does it hold some out?
does nlclassifier perform random sampling to the training datasets?
----
Where can I find documentation on the Natural Language Classifier API?
Where can I find the API documentation?
What is the API for the Classifier?
Where can I find documentation on the NL classifier API?
Where can I find REST API documentation for the NL classifier?
Where can I view documentation for the classifier API?
How do I access the classifier API?

Here "----" implies new class.

Thanks in advance.

from cnn-text-classification-tf.

Hey,@dennybritz
Thanks for taking so much pains.I finally have a working model.
One last query how can we use word2vec model in this?

Thanks & Regards,
Raj Miglani.

from cnn-text-classification-tf.

@rajmiglani See #10 and like zhang-jian said, make sure your vocabulary matches up.

from cnn-text-classification-tf.

@areebak There's an unsubscribe button on the right side. I can't do anything about it.

from cnn-text-classification-tf.

Hey,
Sorry for the troubles guys but as my prev comment mentions i have solved the issues.

from cnn-text-classification-tf.

@zishell
Could you share the full script to predict one sentence? I've tried to use the eval.py, but failed to predict correct. Many thanks. :)

from cnn-text-classification-tf.

@dennybritz you should have a blog post or a video on your YouTube channel just explaining the process to use eval.py or maybe the entire process, which would be very helpful for many of us especially the newbies (including me). Thanks 👍 😃

from cnn-text-classification-tf.

Okay, I refactored the code to make eval.py easier to use. You can now just load any string text data and don't need to deal with the vocabulary generation yourself. The vocabulary is saved during training and then automatically loaded during test. Note that you need to retrain for it to work.

See 9ba22d2 and the new eval.py for an example.

Also, I'm gonna close this for now.

from cnn-text-classification-tf.

If I run the standard model on the Rotten Tomatoes 5331/5331. Is there a way to run the eval.py on only the dev instances from the training? Or do I need to look at the summaries for that.

from cnn-text-classification-tf.

I changed the my own test data like this x_raw = ["./data/metatest.txt"]. But I found that in the output csv only evaluated my training data. How would I change to make eval.py could evaluate my own data?

from cnn-text-classification-tf.