applenob

import tensorflow as tf

def masked_softmax(logits, mask, axis):
    """softmax with mask after the exp operation"""
    e_logits = tf.exp(logits)
    masked_e = tf.multiply(e_logits, mask)
    sum_masked_e = tf.reduce_sum(masked_e, axis, keep_dims=True)
    # be careful when the sum is 0
    ones = tf.ones_like(sum_masked_e)
    sum_masked_e_safe = tf.where(tf.equal(sum_masked_e, 0), ones, sum_masked_e)

applenob

class TopQueue:
    """keep top n num in a list with ascend order."""
    def __init__(self, n):
        self.n = n
        self.queue = []

    def add(self, new_val):
        bigger_index = -1
        for i, one in enumerate(self.queue):
            if one > new_val:

applenob

"""Reference: https://stackoverflow.com/questions/43547402/how-to-calculate-f1-macro-in-keras"""
from keras import backend as K

def f1(y_true, y_pred):
    def recall(y_true, y_pred):
        """Recall metric.

        Only computes a batch-wise average of recall.

        Computes the recall, a metric for multi-label classification of

applenob

#!/usr/bin/bash

# Download zeromq
# Ref http://zeromq.org/intro:get-the-software
wget https://github.com/zeromq/libzmq/releases/download/v4.2.2/zeromq-4.2.2.tar.gz

# Unpack tarball package
tar xvzf zeromq-4.2.2.tar.gz

# Install dependency

applenob

sudo su

# Java
yum -y install java-1.8.0-openjdk-devel

# Build Esentials (minimal)
yum -y install gcc gcc-c++ kernel-devel make automake autoconf swig git unzip libtool binutils

# Extra Packages for Enterprise Linux (EPEL) (for pip, zeromq3)
yum -y install epel-release

applenob

"""
freeze tensorflow checkpoint file to pb format.
"""

import argparse
import os
import tensorflow as tf
import logging
from tensorflow.python.framework import graph_util

applenob

# coding=utf-8

"""Save a tensorflow model to a pb file."""

# Build the model, then train it or load weights from somewhere else.
# ...

graph = tf.get_default_graph()
input_graph_def = graph.as_graph_def()

applenob

# coding=utf-8
import tensorflow as tf


def get_session():
    """load a new session"""
    config = tf.ConfigProto()
    config.gpu_options.allow_growth = True
    return tf.Session(config=config)

applenob

# Refer to https://www.kaggle.com/jhoward/nb-svm-strong-linear-baseline

import re, string
re_tok = re.compile(f'([{string.punctuation}“”¨«»®´·º½¾¿¡§£₤‘’])')
def tokenize(s): 
  return re_tok.sub(r' \1 ', s).split()

applenob

# variation to https://github.com/ryankiros/skip-thoughts/blob/master/decoding/search.py

def keras_rnn_predict(samples, empty=empty, rnn_model=model, maxlen=maxlen):
    """for every sample, calculate probability for every possible label
    you need to supply your RNN model and maxlen - the length of sequences it can handle
    """
    data = sequence.pad_sequences(samples, maxlen=maxlen, value=empty)
    return rnn_model.predict(data, verbose=0)

def beamsearch(predict=keras_rnn_predict,