이혜연

Delete server.py

1 -import numpy as np
2 -import pickle
3 -import tensorflow as tf
4 -from flask import Flask, jsonify, render_template, request
5 -import model
6 -
7 -# Load in data structures
8 -with open("data/wordList.txt", "rb") as fp:
9 - wordList = pickle.load(fp)
10 -wordList.append('<pad>')
11 -wordList.append('<EOS>')
12 -
13 -# Load in hyperparamters
14 -vocabSize = len(wordList)
15 -batchSize = 24
16 -maxEncoderLength = 15
17 -maxDecoderLength = 15
18 -lstmUnits = 112
19 -numLayersLSTM = 3
20 -
21 -# Create placeholders
22 -encoderInputs = [tf.placeholder(tf.int32, shape=(None,)) for i in range(maxEncoderLength)]
23 -decoderLabels = [tf.placeholder(tf.int32, shape=(None,)) for i in range(maxDecoderLength)]
24 -decoderInputs = [tf.placeholder(tf.int32, shape=(None,)) for i in range(maxDecoderLength)]
25 -feedPrevious = tf.placeholder(tf.bool)
26 -
27 -encoderLSTM = tf.nn.rnn_cell.BasicLSTMCell(lstmUnits, state_is_tuple=True)
28 -#encoderLSTM = tf.nn.rnn_cell.MultiRNNCell([singleCell]*numLayersLSTM, state_is_tuple=True)
29 -decoderOutputs, decoderFinalState = tf.contrib.legacy_seq2seq.embedding_rnn_seq2seq(encoderInputs, decoderInputs, encoderLSTM,
30 - vocabSize, vocabSize, lstmUnits, feed_previous=feedPrevious)
31 -
32 -decoderPrediction = tf.argmax(decoderOutputs, 2)
33 -
34 -# Start session and get graph
35 -sess = tf.Session()
36 -#y, variables = model.getModel(encoderInputs, decoderLabels, decoderInputs, feedPrevious)
37 -
38 -# Load in pretrained model
39 -saver = tf.train.Saver()
40 -saver.restore(sess, tf.train.latest_checkpoint('models'))
41 -zeroVector = np.zeros((1), dtype='int32')
42 -
43 -def pred(inputString):
44 - inputVector = model.getTestInput(inputString, wordList, maxEncoderLength)
45 - feedDict = {encoderInputs[t]: inputVector[t] for t in range(maxEncoderLength)}
46 - feedDict.update({decoderLabels[t]: zeroVector for t in range(maxDecoderLength)})
47 - feedDict.update({decoderInputs[t]: zeroVector for t in range(maxDecoderLength)})
48 - feedDict.update({feedPrevious: True})
49 - ids = (sess.run(decoderPrediction, feed_dict=feedDict))
50 - return model.idsToSentence(ids, wordList)
51 -
52 -# webapp
53 -app = Flask(__name__, template_folder='./')
...\ No newline at end of file ...\ No newline at end of file