From dda75cebdbfa2c9d64038be36d60d7dc040f8e32 Mon Sep 17 00:00:00 2001
From: Hashim Sharif <hsharif3@tyler.cs.illinois.edu>
Date: Thu, 20 Aug 2020 16:52:27 -0500
Subject: [PATCH] Adding VGG16_CIFAR100 to new Benchmark structure
---
llvm/projects/keras/src/vgg16_cifar100.py | 111 ++++++++--------------
1 file changed, 39 insertions(+), 72 deletions(-)
diff --git a/llvm/projects/keras/src/vgg16_cifar100.py b/llvm/projects/keras/src/vgg16_cifar100.py
index 66fe6be669..41d0816ecd 100644
--- a/llvm/projects/keras/src/vgg16_cifar100.py
+++ b/llvm/projects/keras/src/vgg16_cifar100.py
@@ -1,6 +1,7 @@
from __future__ import print_function
import os
+import sys
import keras
from keras.datasets import cifar100
from keras.preprocessing.image import ImageDataGenerator
@@ -12,27 +13,20 @@ import numpy as np
from keras.layers.core import Lambda
from keras import backend as K
from keras import regularizers
-from approxhpvm_translator import translate_to_approxhpvm
-import sys
-from weight_utils import dumpCalibrationData
+from Benchmark import Benchmark
+from frontend.weight_utils import dumpCalibrationData
+from frontend.approxhpvm_translator import translate_to_approxhpvm
-class cifar100vgg:
- def __init__(self,train=True):
- self.num_classes = 100
- self.weight_decay = 0.0005
- self.x_shape = [3,32,32]
+class VGG16_CIFAR100(Benchmark):
- self.model = self.build_model()
- if train:
- self.model = self.train(self.model)
- else:
- self.model.load_weights('cifar100vgg.h5')
+ def buildModel(self):
- def build_model(self):
- # Build the network of vgg for 10 classes with massive dropout and weight decay as described in the paper.
+ # Build the network of vgg for 100 classes
+ self.weight_decay = 0.0005
+ self.x_shape = [3,32,32]
model = Sequential()
weight_decay = self.weight_decay
@@ -118,35 +112,21 @@ class cifar100vgg:
return model
- def normalize(self,X_train,X_test):
- #this function normalize inputs for zero mean and unit variance
- # it is used when training a model.
- # Input: training set and test set
- # Output: normalized training set and test set according to the trianing set statistics.
- mean = np.mean(X_train,axis=(0,1,2,3))
- std = np.std(X_train, axis=(0, 1, 2, 3))
- print(mean)
- print(std)
- X_train = (X_train-mean)/(std+1e-7)
- X_test = (X_test-mean)/(std+1e-7)
- return X_train, X_test
- def normalize_production(self,x):
- #this function is used to normalize instances in production according to saved training set statistics
- # Input: X - a training set
- # Output X - a normalized training set according to normalization constants.
+ def data_preprocess(self):
- #these values produced during first training and are general for the standard cifar10 training set normalization
- mean = 121.936
- std = 68.389
- return (x-mean)/(std+1e-7)
+ (X_train, Y_train), (X_test, Y_test) = cifar100.load_data()
- def predict(self,x,normalize=True,batch_size=50):
- if normalize:
- x = self.normalize_production(x)
- return self.model.predict(x,batch_size)
+ mean = np.mean(X_train,axis=(0,1,2,3))
+ std = np.std(X_train,axis=(0,1,2,3))
+ X_train = (X_train-mean)/(std+1e-7)
+ X_test = (X_test-mean)/(std+1e-7)
- def train(self,model):
+ return X_train, Y_train, X_test, Y_test
+
+
+
+ def trainModel(self,model):
#training parameters
batch_size = 128
@@ -166,12 +146,12 @@ class cifar100vgg:
y_train = keras.utils.to_categorical(y_train, self.num_classes)
y_test = keras.utils.to_categorical(y_test, self.num_classes)
-
+
def lr_scheduler(epoch):
return learning_rate * (0.5 ** (epoch // lr_drop))
+
reduce_lr = keras.callbacks.LearningRateScheduler(lr_scheduler)
-
#data augmentation
datagen = ImageDataGenerator(
featurewise_center=False, # set input mean to 0 over the dataset
@@ -188,56 +168,43 @@ class cifar100vgg:
datagen.fit(x_train)
-
#optimization details
sgd = optimizers.SGD(lr=learning_rate, decay=lr_decay, momentum=0.9, nesterov=True)
model.compile(loss='categorical_crossentropy', optimizer=sgd,metrics=['accuracy'])
-
+
# training process in a for loop with learning rate drop every 25 epoches.
-
historytemp = model.fit_generator(datagen.flow(x_train, y_train,
batch_size=batch_size),
steps_per_epoch=x_train.shape[0] // batch_size,
epochs=maxepoches,
validation_data=(x_test, y_test),callbacks=[reduce_lr],verbose=2)
- model.save_weights('cifar100vgg.h5')
+
+ ##### model.save_weights('cifar100vgg.h5')
return model
-if __name__ == '__main__':
- K.set_image_data_format('channels_first')
- os.environ["CUDA_VISIBLE_DEVICES"] = "1"
- (x_train, y_train), (x_test, y_test) = cifar100.load_data()
- test_labels = y_test
- train_labels = y_train
- x_train = x_train.astype('float32')
- x_test = x_test.astype('float32')
- y_train = keras.utils.to_categorical(y_train, 100)
- y_test = keras.utils.to_categorical(y_test, 100)
- model = cifar100vgg()
+
+if __name__ == "__main__":
- predicted_x = model.predict(x_test)
+
+ os.environ["CUDA_VISIBLE_DEVICES"] = "0"
+ # Changing to NCHW format
+ K.set_image_data_format('channels_first')
- norm_test = model.normalize_production(x_test)
- x_train = model.normalize_production(x_train)
-
- dumpCalibrationData("calibration_data/vgg16_cifar100_calib.bin", x_train,
- "calibration_data/vgg16_cifar100_train_labels.bin", train_labels)
- sys.exit(0)
-
-
- translate_to_approxhpvm(model.model, "vgg16_cifar100_test/", norm_test, test_labels,
- "vgg16_cifar100_front", y_test)
+ ### Parameters specific to each benchmark
+ reload_dir = "/home/hsharif3/Gitlab/hpvm/llvm/projects/hpvm-tensor-rt/model_params/vgg16_cifar100/"
+ keras_model_file = "vgg16_cifar100.h5"
+ hpvm_dir = "data/vgg16_cifar100/"
+ num_classes = 100
+ vgg16_cifar100 = VGG16_CIFAR100("vgg16_cifar100", reload_dir, keras_model_file, hpvm_dir, num_classes)
- residuals = (np.argmax(predicted_x,1)!=np.argmax(y_test,1))
- loss = sum(residuals)/len(residuals)
- print("the validation 0/1 loss is: ",loss)
-
+ vgg16_cifar100.run(sys.argv)
+
--
GitLab