03-mlp_score_on_transfert.py
3.38 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
# coding: utf-8
# In[2]:
# Import
import gensim
from scipy import sparse
import itertools
from sklearn import preprocessing
from keras.models import Sequential
from keras.optimizers import SGD,Adam
from keras.layers.advanced_activations import ELU,PReLU
from keras.callbacks import ModelCheckpoint
from mlp import *
import sklearn.metrics
from sklearn.preprocessing import LabelBinarizer
import shelve
import pickle
from utils import *
import sys
import os
import json
# In[4]:
in_dir = sys.argv[1]
#['ASR', 'TRS', 'LABEL']
# In[6]:
json_conf =json.load(open(sys.argv[2]))
mlp_conf = json_conf["mlp"]
hidden_size = mlp_conf["hidden_size"]
loss = mlp_conf["loss"]
patience = mlp_conf["patience"]
dropouts = mlp_conf["do"]
epochs = mlp_conf["epochs"]
batch_size = mlp_conf["batch"]
input_activation=mlp_conf["input_activation"]
output_activation=mlp_conf["output_activation"]
try:
k = mlp_conf["sgd"]
if mlp_conf["sgd"]["name"] == "adam":
sgd = Adam(lr=mlp_conf["sgd"]["lr"])#SGD(lr=0.00001,nesterov=False) #'rmsprop'# Adam(lr=0.00001)#SGD(lr=0.001, momentum=0.9, nesterov=True)
elif mlp_conf["sgd"]["name"] == "sgd":
sgd = SGD(lr=mlp_conf["sgd"]["lr"])
except:
sgd = mlp_conf["sgd"]
name = json_conf["name"]
db = shelve.open("{}/{}/labels.shelve".format(in_dir,name))
shelve_logs=shelve.open("{}/{}/03_logs.shelve".format(in_dir,name),writeback=True)
#
keys = db["LABEL"].keys()
hdf_proj_path = "{}/{}/transfert_proj_df.hdf".format(in_dir,name)
proj_hdf = pandas.HDFStore(hdf_proj_path)
hdf_keys = proj_hdf.keys()
print hdf_keys
proj_hdf.close()
hdf_lvl = set([ x.split("/")[1] for x in hdf_keys ])
hdf_layer = set( [ x.split("/")[2] for x in hdf_keys ])
hdf_crossval = set([ x.split("/")[3] for x in hdf_keys ])
print hdf_lvl
print hdf_crossval
labels_dict = { }
logs = {}
for lvl in hdf_lvl :
labels_dict[lvl] = {}
for layer in hdf_layer:
labels_dict[lvl][layer] = {}
for lvl in hdf_lvl :
for layer in hdf_layer:
x_train = pandas.read_hdf(hdf_proj_path,key="/{}/{}/{}".format(lvl,layer,"TRAIN"))
x_dev = pandas.read_hdf(hdf_proj_path,key="/{}/{}/{}".format(lvl,layer,"DEV"))
x_test = pandas.read_hdf(hdf_proj_path,key="/{}/{}/{}".format(lvl,layer, "TEST"))
y_train = db["LABEL"]["ASR"]["TRAIN"]
y_dev = db["LABEL"]["ASR"]["DEV"]
y_test = db["LABEL"]["ASR"]["TEST"]
print x_train.shape
print x_dev.shape
print x_test.shape
print y_train.shape
print y_dev.shape
print y_test.shape
pred,hist = train_mlp_pred(x_train.values,y_train,
x_dev.values,y_dev,
x_test.values,y_test,
hidden_size ,sgd=sgd,
epochs=epochs,
patience=patience,
batch_size=batch_size,
input_activation=input_activation,
output_activation=output_activation,
dropouts=dropouts,
fit_verbose=1)
shelve_logs["{}/{}".format(lvl,layer)] = hist
labels_dict[lvl][layer]["TRAIN"] = np.argmax(pred[0],axis=1)
labels_dict[lvl][layer]["DEV"] = np.argmax(pred[1],axis=1)
labels_dict[lvl][layer]["TEST"] = np.argmax(pred[2],axis=1)
db["transfert"] = labels_dict
shelve_logs.sync()
shelve_logs.close()
db.sync()
db.close()