Commit eb14f11a authored by Malsha Rathnasiri's avatar Malsha Rathnasiri

final model

parent 40d2aa63
...@@ -3,4 +3,4 @@ from django.apps import AppConfig ...@@ -3,4 +3,4 @@ from django.apps import AppConfig
class CmsConfig(AppConfig): class CmsConfig(AppConfig):
default_auto_field = 'django.db.models.BigAutoField' default_auto_field = 'django.db.models.BigAutoField'
name = 'cms' name = 'backend.cms'
# Generated by Django 4.0.4 on 2022-04-25 19:51
from django.db import migrations, models
class Migration(migrations.Migration):
initial = True
dependencies = [
]
operations = [
migrations.CreateModel(
name='MlModel',
fields=[
('id', models.AutoField(primary_key=True, serialize=False)),
('description', models.TextField(blank=True)),
('timestamp', models.DateTimeField(auto_now=True)),
('details', models.JSONField()),
],
),
]
import pickle
from keras.models import load_model
import numpy as np
import IPython.display as ipd
import random
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
def predict():
model=load_model(r'./best_model_final.hdf5')
f1 = open('all_label.txt', 'rb')
all_label = pickle.load(f1)
print('loaded labels')
f2 = open('all_waves_file.txt', 'rb')
all_wave = pickle.load(f2)
print('loaded waves')
le = LabelEncoder()
y = le.fit_transform(all_label)
classes = list(le.classes_)
train_data_file = open("train_data_file.txt", 'rb')
[x_tr, x_val, y_tr, y_val] = np.load(train_data_file, allow_pickle=True)
train_data_file.close()
def predictSamples(audio):
prob=model.predict(audio.reshape(1,8000,1))
index=np.argmax(prob[0])
return classes[index]
index=random.randint(0,len(x_val)-1)
samples=x_val[index].ravel()
print("Audio:",classes[np.argmax(y_val[index])])
ipd.Audio(samples, rate=8000)
print("Text:",predictSamples(samples))
\ No newline at end of file
import pickle
from matplotlib import pyplot
import os import os
import librosa import librosa
import IPython.display as ipd import IPython.display as ipd
...@@ -6,84 +8,192 @@ import numpy as np ...@@ -6,84 +8,192 @@ import numpy as np
from scipy.io import wavfile from scipy.io import wavfile
import warnings import warnings
from sklearn.preprocessing import LabelEncoder from sklearn.preprocessing import LabelEncoder
from keras.utils import np_utils from keras.utils import np_utils
warnings.filterwarnings("ignore") from sklearn.model_selection import train_test_split
os.listdir('../input/') from keras.layers import Dense, Dropout, Flatten, Conv1D, Input, MaxPooling1D
from keras.models import Model
from keras.callbacks import EarlyStopping, ModelCheckpoint
from keras import backend as K
K.clear_session()
train_audio_path = '../input/tensorflow-speech-recognition-challenge/train/audio/' warnings.filterwarnings("ignore")
samples, sample_rate = librosa.load(train_audio_path+'yes/0a7c2a8d_nohash_0.wav', sr = 16000)
fig = plt.figure(figsize=(14, 8))
ax1 = fig.add_subplot(211)
ax1.set_title('Raw wave of ' + '../input/train/audio/yes/0a7c2a8d_nohash_0.wav')
ax1.set_xlabel('time')
ax1.set_ylabel('Amplitude')
ax1.plot(np.linspace(0, sample_rate/len(samples), sample_rate), samples)
# os.listdir('../../../data/')
ipd.Audio(samples, rate=sample_rate) classes = ['down', 'go', 'left', 'no', 'off',
'on', 'right', 'stop', 'up', 'yes']
print(sample_rate)
def train():
samples = librosa.resample(samples, sample_rate, 8000) print('1')
ipd.Audio(samples, rate=8000) train_audio_path = r'./backend/data/train/train/audio/'
samples, sample_rate = librosa.load(
train_audio_path+'yes/0a7c2a8d_nohash_0.wav', sr=16000)
# fig = plt.figure(figsize=(14, 8))
# ax1 = fig.add_subplot(211)
# ax1.set_title('Raw wave of ' + r'../input/train/audio/yes/0a7c2a8d_nohash_0.wav')
# ax1.set_xlabel('time')
# ax1.set_ylabel('Amplitude')
# ax1.plot(np.linspace(0, sample_rate/len(samples), sample_rate), samples)
ipd.Audio(samples, rate=sample_rate)
labels=os.listdir(train_audio_path) print(sample_rate)
#find count of each label and plot bar graph samples = librosa.resample(samples, sample_rate, 8000)
no_of_recordings=[] ipd.Audio(samples, rate=8000)
for label in labels:
waves = [f for f in os.listdir(train_audio_path + '/'+ label) if f.endswith('.wav')] labels = os.listdir(train_audio_path)
# find count of each label and plot bar graph
no_of_recordings = []
for label in labels:
waves = [f for f in os.listdir(
train_audio_path + '/' + label) if f.endswith('.wav')]
no_of_recordings.append(len(waves)) no_of_recordings.append(len(waves))
#plot # plot
plt.figure(figsize=(30,5)) # plt.figure(figsize=(30,5))
index = np.arange(len(labels)) index = np.arange(len(labels))
plt.bar(index, no_of_recordings) # plt.bar(index, no_of_recordings)
plt.xlabel('Commands', fontsize=12) # plt.xlabel('Commands', fontsize=12)
plt.ylabel('No of recordings', fontsize=12) # plt.ylabel('No of recordings', fontsize=12)
plt.xticks(index, labels, fontsize=15, rotation=60) # plt.xticks(index, labels, fontsize=15, rotation=60)
plt.title('No. of recordings for each command') # plt.title('No. of recordings for each command')
plt.show() # plt.show()
labels=["yes", "no", "up", "down", "left", "right", "on", "off", "stop", "go"] print('2')
duration_of_recordings=[] labels = ["yes", "no", "up", "down", "left",
for label in labels: "right", "on", "off", "stop", "go"]
waves = [f for f in os.listdir(train_audio_path + '/'+ label) if f.endswith('.wav')]
# labels_file = open('./labels_file.bin', 'wb+')
# pickle.dump(obj=labels, file=labels_file)
# labels_file.close()
# # file = open('./labels_file.bin', 'rb')
# # dict = pickle.load(file)
# # print('loaded')
# # print(dict)
# # print('fdnasf')
duration_of_recordings = []
for label in labels:
print('2.1', label)
waves = [f for f in os.listdir(
train_audio_path + '/' + label) if f.endswith('.wav')]
for wav in waves: for wav in waves:
sample_rate, samples = wavfile.read(train_audio_path + '/' + label + '/' + wav) sample_rate, samples = wavfile.read(
train_audio_path + '/' + label + '/' + wav)
duration_of_recordings.append(float(len(samples)/sample_rate)) duration_of_recordings.append(float(len(samples)/sample_rate))
plt.hist(np.array(duration_of_recordings)) plt.hist(np.array(duration_of_recordings))
train_audio_path = '../input/tensorflow-speech-recognition-challenge/train/audio/' train_audio_path = r'./backend/data/train/train/audio/'
all_wave = [] # all_wave = []
all_label = [] # all_label = []
for label in labels: # for label in labels:
print(label) # print(label)
waves = [f for f in os.listdir(train_audio_path + '/'+ label) if f.endswith('.wav')] # waves = [f for f in os.listdir(
for wav in waves: # train_audio_path + '/' + label) if f.endswith('.wav')]
samples, sample_rate = librosa.load(train_audio_path + '/' + label + '/' + wav, sr = 16000) # for wav in waves:
samples = librosa.resample(samples, sample_rate, 8000) # samples, sample_rate = librosa.load(
if(len(samples)== 8000) : # train_audio_path + '/' + label + '/' + wav, sr=16000)
all_wave.append(samples) # samples = librosa.resample(samples, sample_rate, 8000)
all_label.append(label) # if(len(samples) == 8000):
# all_wave.append(samples)
# all_label.append(label)
# print('3')
f1 = open('all_label.txt', 'rb')
all_label = pickle.load(f1)
print('loaded labels')
f2 = open('all_waves_file.txt', 'rb')
all_wave = pickle.load(f2)
print('loaded waves')
le = LabelEncoder()
y = le.fit_transform(all_label)
classes = list(le.classes_)
print(classes)
print(all_wave)
print('4')
y = np_utils.to_categorical(y, num_classes=len(labels))
all_wave = np.array(all_wave).reshape(-1, 8000, 1)
x_tr, x_val, y_tr, y_val = train_test_split(np.array(all_wave), np.array(
y), stratify=y, test_size=0.2, random_state=777, shuffle=True)
train_data_file = open('train_data_file.txt', 'wb+')
np.save(file=train_data_file, arr=np.array([x_tr, x_val, y_tr, y_val]))
train_data_file.close()
inputs = Input(shape=(8000, 1))
# First Conv1D layer
conv = Conv1D(8, 13, padding='valid', activation='relu', strides=1)(inputs)
conv = MaxPooling1D(3)(conv)
conv = Dropout(0.3)(conv)
# Second Conv1D layer
conv = Conv1D(16, 11, padding='valid', activation='relu', strides=1)(conv)
conv = MaxPooling1D(3)(conv)
conv = Dropout(0.3)(conv)
# Third Conv1D layer
conv = Conv1D(32, 9, padding='valid', activation='relu', strides=1)(conv)
conv = MaxPooling1D(3)(conv)
conv = Dropout(0.3)(conv)
# Fourth Conv1D layer
conv = Conv1D(64, 7, padding='valid', activation='relu', strides=1)(conv)
conv = MaxPooling1D(3)(conv)
conv = Dropout(0.3)(conv)
# Flatten layer
conv = Flatten()(conv)
# Dense Layer 1
conv = Dense(256, activation='relu')(conv)
conv = Dropout(0.3)(conv)
# Dense Layer 2
conv = Dense(128, activation='relu')(conv)
conv = Dropout(0.3)(conv)
outputs = Dense(len(labels), activation='softmax')(conv)
model = Model(inputs, outputs)
model.summary()
model.compile(loss='categorical_crossentropy',
optimizer='adam', metrics=['accuracy'])
es = EarlyStopping(monitor='val_loss', mode='min',
verbose=1, patience=10, min_delta=0.0001)
mc = ModelCheckpoint('best_model.hdf5', monitor='val_accuracy',
verbose=1, save_best_only=True, mode='max')
history = model.fit(x_tr, y_tr, epochs=100, callbacks=[
es, mc], batch_size=32, validation_data=(x_val, y_val))
le = LabelEncoder() # pyplot.plot(history.history['loss'], label='train')
y=le.fit_transform(all_label) # pyplot.plot(history.history['val_loss'], label='test')
classes= list(le.classes_) # pyplot.legend()
# pyplot.show()
y=np_utils.to_categorical(y, num_classes=len(labels)) return history
\ No newline at end of file
from django.db import models from django.db import models
# Create your models here. # Create your models here.
class MlModel(models.Model):
id = models.AutoField(primary_key=True)
description = models.TextField(blank=True)
timestamp = models.DateTimeField(blank=True, auto_now=True)
details = models.JSONField()
from django.contrib.auth.models import User, Group from django.contrib.auth.models import User, Group
from .models import MlModel
from rest_framework import serializers from rest_framework import serializers
...@@ -12,3 +13,10 @@ class GroupSerializer(serializers.HyperlinkedModelSerializer): ...@@ -12,3 +13,10 @@ class GroupSerializer(serializers.HyperlinkedModelSerializer):
class Meta: class Meta:
model = Group model = Group
fields = ['url', 'name'] fields = ['url', 'name']
class MlModelSerializer(serializers.ModelSerializer):
class Meta:
model = MlModel
fields = (
'__all__'
)
\ No newline at end of file
...@@ -2,10 +2,17 @@ from http.client import HTTPResponse ...@@ -2,10 +2,17 @@ from http.client import HTTPResponse
from django.contrib.auth.models import User, Group from django.contrib.auth.models import User, Group
from rest_framework import viewsets from rest_framework import viewsets
from rest_framework import permissions from rest_framework import permissions
from backend.cms.serializers import MlModelSerializer
from backend.cms.serializers import UserSerializer, GroupSerializer from backend.cms.serializers import UserSerializer, GroupSerializer
from rest_framework.decorators import action from rest_framework.decorators import action
from rest_framework.response import Response from rest_framework.response import Response
from .models import MlModel
from .model.train import train
from .model.predict import predict
class UserViewSet(viewsets.ModelViewSet): class UserViewSet(viewsets.ModelViewSet):
""" """
...@@ -20,6 +27,9 @@ class UserViewSet(viewsets.ModelViewSet): ...@@ -20,6 +27,9 @@ class UserViewSet(viewsets.ModelViewSet):
print('Function ran') print('Function ran')
results = train()
print(results)
return Response({'success': True}) return Response({'success': True})
...@@ -30,3 +40,28 @@ class GroupViewSet(viewsets.ModelViewSet): ...@@ -30,3 +40,28 @@ class GroupViewSet(viewsets.ModelViewSet):
queryset = Group.objects.all() queryset = Group.objects.all()
serializer_class = GroupSerializer serializer_class = GroupSerializer
permission_classes = [permissions.IsAuthenticated] permission_classes = [permissions.IsAuthenticated]
class MlModelViewSet(viewsets.ViewSet):
queryset = MlModel.objects.all()
serializer_class = MlModelSerializer
permission_classes = [permissions.IsAuthenticated]
@action(detail=False)
def train(*args, **kwargs):
print('Function ran')
results = train()
print(results)
return Response({'success': True})
@action(detail=False)
def predict(*args, **kwargs):
print('Function ran')
results = predict()
print(results)
return Response({'success': True})
...@@ -38,6 +38,7 @@ INSTALLED_APPS = [ ...@@ -38,6 +38,7 @@ INSTALLED_APPS = [
'django.contrib.messages', 'django.contrib.messages',
'django.contrib.staticfiles', 'django.contrib.staticfiles',
'rest_framework', 'rest_framework',
'backend.cms'
] ]
MIDDLEWARE = [ MIDDLEWARE = [
......
...@@ -5,6 +5,7 @@ from backend.cms import views ...@@ -5,6 +5,7 @@ from backend.cms import views
router = routers.DefaultRouter() router = routers.DefaultRouter()
router.register(r'users', views.UserViewSet) router.register(r'users', views.UserViewSet)
router.register(r'groups', views.GroupViewSet) router.register(r'groups', views.GroupViewSet)
router.register(r'mlmodels', views.MlModelViewSet)
# Wire up our API using automatic URL routing. # Wire up our API using automatic URL routing.
# Additionally, we include login URLs for the browsable API. # Additionally, we include login URLs for the browsable API.
......
No preview for this file type
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment