re4388

!pip3 install --quiet "tensorflow>=1.7"
!pip3 install --quiet tensorflow-hub
!pip3 install seaborn
import tensorflow as tf
import tensorflow_hub as hub
import matplotlib.pyplot as plt
import numpy as np
import os
import pandas as pd
import re

re4388

['HUM', 'ENTY', 'LOC', 'ABBR', 'DESC', 'NUM']

re4388

new_text = ["current us president?", "what compose my pc?", 'where is USA', 'what U.S.A mean?', 
'tell me about how economy work', 'the number of PI?']

re4388

import seaborn as sns
import matplotlib.pyplot as plt


df_train = get_dataframe('test_data.txt')
category_counts = len(df_train.label.cat.categories)
print(category_counts) # 6

sns.barplot(x.index,x)
plt.gca().set_ylabel('samples')

re4388

def UniversalEmbedding(x):
    return embed(tf.squeeze(tf.cast(x, tf.string)), signature="default", as_dict=True)["default"]

input_text = layers.Input(shape=(1,), dtype=tf.string)
embedding = layers.Lambda(UniversalEmbedding, output_shape=(embed_size,))(input_text)
dense = layers.Dense(256, activation='relu')(embedding)
pred = layers.Dense(category_counts, activation='softmax')(dense)

model = Model(inputs=[input_text], outputs=pred)

re4388

train_text = np.array(df_train['text'].tolist(), dtype=object)[:, np.newaxis]
train_label = np.asarray(pd.get_dummies(df_train.label), dtype = np.int8)

print(train_text.shape)
print(train_label.shape)
print(train_label[:3])

# also handel the test set
df_test = get_dataframe('test_data.txt')
test_text = np.array(df_test['text'].tolist(), dtype=object)[:, np.newaxis]

re4388

with tf.Session() as session:
  K.set_session(session)
  session.run(tf.global_variables_initializer())
  session.run(tf.tables_initializer())

  history = model.fit(train_text, train_label,
            validation_data=(test_text, test_label),
            epochs=10,
            batch_size=32)
  model.save_weights('./model.h5')

re4388

def run_pred(text_list):
    with tf.Session() as session:
        K.set_session(session)
        session.run(tf.global_variables_initializer())
        session.run(tf.tables_initializer())
        model.load_weights('./model.h5')  
        predicts = model.predict(np.array(text_list, dtype=object)[:, np.newaxis], batch_size=32)
        categories = df_train.label.cat.categories.tolist()
    predict_labels = [categories[logit] for logit in predicts.argmax(axis=1)]
    return predict_labels

re4388

# code origin: https://www.jianshu.com/p/e1d95c4e1697

class Entity:
def __init__(self, size, x, y):
    """ init x, y when init the instance"""
    self.x, self.y = x, y
    self.size = size

def __call__(self, x, y):
    '''you can change x, y after'''

re4388

# code origin: https://www.kaggle.com/timesler/facial-recognition-model-in-pytorch

class DetectionPipeline:
    """Pipeline class for detecting faces in the frames of a video file."""
    
    def __init__(self, detector, n_frames=None, batch_size=60, resize=None):
        """Constructor for DetectionPipeline class.
        
        Keyword Arguments:
            n_frames {int} -- Total number of frames to load. These will be evenly spaced