dalequark

import plotly.plotly as py
# (*) Useful Python/Plotly tools
import plotly.tools as tls
from plotly.graph_objs import *
import numpy as np
import random
import sys, serial
import urllib2

dalequark

# We'll consider everything with a Normalized Score below the 25th percentile to be rude.
# Above the 75th percentile is considered polite.
# Scores in the middle are considered neutral

# Get the 25th and 75th percentiles of Normalized Scoes
rude_thresh = data.describe().loc["25%"]['Normalized Score']
polite_thresh = data.describe().loc["75%"]['Normalized Score']

label_list = ["rude", "neutral", "polite"]
def score_to_label(score):

dalequark

from sklearn.model_selection import train_test_split
import pandas as pd
import tensorflow as tf
import tensorflow_hub as hub
from datetime import datetime

# Install and import BERT's libraries
!pip install bert-tensorflow
import bert
from bert import run_classifier

dalequark

from tensorflow import keras
import os
import re

# Load all files from a directory in a DataFrame.
def load_directory_data(directory):
  data = {}
  data["sentence"] = []
  data["sentiment"] = []
  for file_path in os.listdir(directory):

dalequark

# Use the InputExample class from BERT's run_classifier code to create examples from the data
train_InputExamples = train.apply(lambda x: bert.run_classifier.InputExample(guid=None, # Globally unique ID for bookkeeping, unused in this example
                                                                   text_a = x[DATA_COLUMN], 
                                                                   text_b = None, 
                                                                   label = x[LABEL_COLUMN]), axis = 1)

test_InputExamples = test.apply(lambda x: bert.run_classifier.InputExample(guid=None, 
                                                                   text_a = x[DATA_COLUMN], 
                                                                   text_b = None, 
                                                                   label = x[LABEL_COLUMN]), axis = 1)

dalequark

# This is a path to an uncased (all lowercase) version of BERT
BERT_MODEL_HUB = "https://tfhub.dev/google/bert_uncased_L-12_H-768_A-12/1"

def create_tokenizer_from_hub_module():
  """Get the vocab file and casing info from the Hub module."""
  with tf.Graph().as_default():
    bert_module = hub.Module(BERT_MODEL_HUB)
    tokenization_info = bert_module(signature="tokenization_info", as_dict=True)
    with tf.Session() as sess:
      vocab_file, do_lower_case = sess.run([tokenization_info["vocab_file"],

dalequark

def create_model(is_predicting, input_ids, input_mask, segment_ids, labels,
                 num_labels):
  """Creates a classification model."""

  bert_module = hub.Module(
      BERT_MODEL_HUB,
      trainable=True)
  bert_inputs = dict(
      input_ids=input_ids,
      input_mask=input_mask,

dalequark

# model_fn_builder actually creates our model function
# using the passed parameters for num_labels, learning_rate, etc.
def model_fn_builder(num_labels, learning_rate, num_train_steps,
                     num_warmup_steps):
  """Returns `model_fn` closure for TPUEstimator."""
  def model_fn(features, labels, mode, params):  # pylint: disable=unused-argument
    """The `model_fn` for TPUEstimator."""

    input_ids = features["input_ids"]
    input_mask = features["input_mask"]

dalequark

# Compute train and warmup steps from batch size
# These hyperparameters are copied from this colab notebook (https://colab.sandbox.google.com/github/tensorflow/tpu/blob/master/tools/colab/bert_finetuning_with_cloud_tpus.ipynb)
BATCH_SIZE = 32
LEARNING_RATE = 2e-5
NUM_TRAIN_EPOCHS = 3.0
# Warmup is a period of time where hte learning rate 
# is small and gradually increases--usually helps training.
WARMUP_PROPORTION = 0.1
# Model configs
SAVE_CHECKPOINTS_STEPS = 500

dalequark

# Load our test data
test_input_fn = run_classifier.input_fn_builder(
    features=test_features,
    seq_length=MAX_SEQ_LENGTH,
    is_training=False,
    drop_remainder=False)
    
    
estimator.evaluate(input_fn=test_input_fn, steps=None)