Commit 2434380a authored by Rudolf Chrispens's avatar Rudolf Chrispens
Browse files

starting to change the None part

parent 9e613e34

.gitignore

+7 −1
Original line number Diff line number Diff line
@@ -4,13 +4,19 @@ cnn_exercise/ 
anaconda3/



#lstm worked files

caption-lib/lstm/Trained_Data

capiton-lib/lstm/Input_Data

capiton-lib/lstm/Read_Data

capiton-lib/log



lstm/Trained_Data

caption-lib/lstm/Trained_Data

log





lstm/Trained_Models/*

capiton-lib/lstm/Trained_Data/*

caption-lib/.vscode/settings.json



.vscode/settings.json

*.pickle

caption-lib/lstm/Models/coco_model.py

+4 −2
Original line number Diff line number Diff line
@@ -52,8 +52,10 @@ class coco_model(object): 
        self.emb_initializer = tf.random_uniform_initializer(minval=-1.0, maxval=1.0)



        # Place holder for features and captions

        self.features = tf.placeholder(tf.float32, [None, self.L, self.D])

        self.captions = tf.placeholder(tf.int32, [None, self.T + 1])

        self.features = tf.placeholder(tf.float32, [self.L, self.D], name="features_placeholder_model")

        self.captions = tf.placeholder(tf.int32, [self.T + 1], name="captions_placeholder_model")

        #self.features = tf.placeholder(tf.float32, [self.L, self.D], name="features_placeholder_model")

        #self.captions = tf.placeholder(tf.int32, [self.T + 1], name="captions_placeholder_model")



    def _get_initial_lstm(self, features):

        with tf.variable_scope('initial_lstm'):

caption-lib/lstm/Readers/coco_reader.py

+27 −12
Original line number Diff line number Diff line
@@ -27,31 +27,46 @@ class coco_reader: 
            for x in d['captions']:

                formatted_data.append([x['caption'], d['file_name'], x['image_id']])

        """



        for data_from_image in data:

            features.append(data_from_image['vector'])

            captions.append(data_from_image['captions'][0]['caption'])

            captions.append(data_from_image['captions'][1]['caption'])

            captions.append(data_from_image['captions'][2]['caption'])

            captions.append(data_from_image['captions'][3]['caption'])

            captions.append(data_from_image['captions'][4]['caption'])

            #captions.append(data_from_image['captions'][1]['caption'])

            #captions.append(data_from_image['captions'][2]['caption'])

            #captions.append(data_from_image['captions'][3]['caption'])

            #captions.append(data_from_image['captions'][4]['caption'])

            image_idxs.append(data_from_image['id'])



            if(config.reader.verbose):

                print('#> reader printing example data:')

                print('\t\tcaptions:', data[0]['captions'][0]['caption'])

                print('\t\tcaptions:', data[0]['captions'][1]['caption'])

                print('\t\tcaptions:', data[0]['captions'][2]['caption'])

                print('\t\tcaptions:', data[0]['captions'][3]['caption'])

                print('\t\tcaptions:', data[0]['captions'][4]['caption'])

                print('\tvector:', data[0]['vector'])

        formatted_data['features'] = np.array(features)

                print('\t\tcaptions:', data_from_image['captions'][0]['caption'])

                #print('\t\tcaptions:', data[0]['captions'][1]['caption'])

                #print('\t\tcaptions:', data[0]['captions'][2]['caption'])

                #print('\t\tcaptions:', data[0]['captions'][3]['caption'])

                #print('\t\tcaptions:', data[0]['captions'][4]['caption'])

                print('\timage_id:', data_from_image['id'])

                print('\tvector:', data_from_image['vector'])



        formatted_data['captions'] = np.array(captions)

        formatted_data['image_idxs'] = np.array(image_idxs)

        formatted_data['features'] = np.array(features)



        end_t = time.time()

        print("#> Load data elapsed time: %.2f" % (end_t - start_t))

        return formatted_data

    """

        def _build_indexed_array(self, masa_data, array):

            nd_array = np.ndarray(len(masa_data), dtype=np.int32)

            for i, value in enumerate(array):

                nd_array[i] = value

            return nd_array



        def _build_indexed_array_features(self, masa_data, array):

            nd_array = np.ndarray((len(masa_data), len(array)), dtype=np.int32)

            for i, value in enumerate(array):

                for x, t in enumerate(value):

                    nd_array[i][x] = t

            return nd_array

    """



    def load_data_masa(self, data_path=helper.inputPath, filename=config.path.input_train):

        #this is the code to load our data like we planned

caption-lib/lstm/Trainers/coco_trainer.py

+24 −6
Original line number Diff line number Diff line
@@ -123,15 +123,33 @@ class coco_trainer(object): 


            for e in range(self.n_epochs):

                rand_idxs = np.random.permutation(n_examples)

                captions = captions[rand_idxs]

                image_idxs = image_idxs[rand_idxs]

                m_captions = captions[rand_idxs]

                m_image_idxs = image_idxs[rand_idxs]

                m_features = features[rand_idxs]



                for i in range(n_iters_per_epoch):

                    captions_batch = captions[i*self.batch_size:(i+1)*self.batch_size]

                    image_idxs_batch = image_idxs[i*self.batch_size:(i+1)*self.batch_size]

                    features_batch = features[image_idxs_batch]

                    captions_batch = m_captions[i*self.batch_size:(i+1)*self.batch_size]

                    image_idxs_batch = m_image_idxs[i*self.batch_size:(i+1)*self.batch_size]

                    features_batch = m_features[i*self.batch_size:(i+1)*self.batch_size]



                    print()

                    print()

                    print()

                    print()

                    print()

                    print()

                    print("image", features_batch[0])

                    print("image_idx", image_idxs_batch[0])

                    print("caption", captions_batch[0])



                    #Here was an ERROR

                    feed_dict = {self.model.features: features_batch, self.model.captions: captions_batch}

                    print()

                    print()

                    print("Feed Dict:")

                    print(feed_dict)

                    print()

                    print()

                    _, l = sess.run([train_op, loss], feed_dict)

                    curr_loss += l
@@ -142,7 +160,7 @@ class coco_trainer(object): 


                    if (i+1) % self.print_every == 0:

                        print("\nTrain loss at epoch %d & iteration %d (mini-batch): %.5f" %(e+1, i+1, l))

                        ground_truths = captions[image_idxs == image_idxs_batch[0]]

                        ground_truths = m_captions[image_idxs == image_idxs_batch[0]]

                        decoded = decode_captions(ground_truths, self.model.idx_to_word)

                        for j, gt in enumerate(decoded):

                            print("Ground truth %d: %s" %(j+1, gt))

caption-lib/lstm/lstm_main.py

+1 −1
Original line number Diff line number Diff line
@@ -38,7 +38,7 @@ def main(self, parameter_list): 
    val_data = current_reader.load_data(filename=_config.path.input_validate)



    model = _model.coco_model(  word_to_idx,

                                dim_feature=[196, 512],

                                dim_feature=[128, 2048],

                                dim_embed=512,

                                dim_hidden=1024,

                                n_time_step=16,