Ejemplos de BasicNormalizer en Python

Ejemplo n.º 1

Archivo: context.py Proyecto: nayutaya/20160228-gdg-kobe

import tensorflow as tf

import normalizer as no
import tokenizer as to
import mlp

categories = ["rail", "other"]
num_of_steps = 500
tokenizer = to.NormalizedTokenizer(normalizer=no.BasicNormalizer(),
                                   tokenizer=to.LetterBigramTokenizer())


def make_model(num_of_terms):
    optimizer = tf.train.AdamOptimizer(0.01)
    model = mlp.MultiLayerPerceptron(optimizer=optimizer,
                                     categories=categories,
                                     num_of_terms=num_of_terms,
                                     num_of_hidden_nodes=100)
    print("model.optimizer           = " + str(type(model.optimizer)))
    print("model.num_of_categories   = " + str(model.num_of_categories))
    print("model.num_of_terms        = " + str(model.num_of_terms))
    print("model.num_of_hidden_nodes = " + str(model.num_of_hidden_nodes))
    return model

Ejemplo n.º 2

import tensorflow as tf

import normalizer as no
import tokenizer as to
import sr

categories   = ["rail", "other"]
num_of_steps = 500
tokenizer    = to.NormalizedTokenizer(
    normalizer=no.BasicNormalizer(),
    tokenizer=to.JanomeBigramTokenizer())

def make_model(num_of_terms):
    optimizer = tf.train.GradientDescentOptimizer(0.01)
    model = sr.SoftmaxRegressions(optimizer=optimizer, categories=categories, num_of_terms=num_of_terms)
    print("model.optimizer         = " + str(type(model.optimizer)))
    print("model.num_of_categories = " + str(model.num_of_categories))
    print("model.num_of_terms      = " + str(model.num_of_terms))
    return model