AlphaZero version 5¶
This model was trained from scratch on various StageTwo Datasets for two epochs each.
Take a look at alphaZeroV5.py and alphaZeroV5.training.log for the correct version. It didn't work using this jupyter notebook, because freeing up memory didn't work.
In [1]:
import sys
sys.path.append('..')
import numpy as np
import tensorflow as tf
from tensorflow.python import debug as tf_debug
from keras.callbacks import *
from keras.models import *
from keras.layers import *
from keras.optimizers import *
from keras.utils.np_utils import to_categorical
from keras.utils import plot_model
import keras.backend as K
from keras.regularizers import l2
from keras.engine.topology import Layer
from PIL import Image
from matplotlib.pyplot import imshow
%matplotlib inline
import random
import gc
from LineFilterLayer import LineFilterLayer
modelPath = 'model/alphaZeroV5.h5'
datasetList = [
'StageTwo-1000-5x4-15:53-09_04_2018.npz',
'StageTwo-1000000-6x5-23:21-08_04_2018.npz',
'StageTwo-1000000-6x5-23:41-08_04_2018.npz',
'StageTwo-1000000-6x5-08:32-09_04_2018.npz',
'StageTwo-1000000-6x5-08:51-09_04_2018.npz',
'StageTwo-1000000-6x5-09:10-09_04_2018.npz',
'StageTwo-1000000-6x5-09:28-09_04_2018.npz',
'StageTwo-1000000-6x5-09:47-09_04_2018.npz',
'StageTwo-1000000-6x5-10:06-09_04_2018.npz',
'StageTwo-1000000-6x5-10:25-09_04_2018.npz',
'StageTwo-1000000-6x5-10:44-09_04_2018.npz',
'StageTwo-1000000-6x5-11:03-09_04_2018.npz',
'StageTwo-1000000-6x5-11:21-09_04_2018.npz',
]
In [2]:
print(K.image_data_format())
# expected output: channels_last
assert(K.image_data_format() == 'channels_last' )
In [5]:
def dotsAndBoxesToCategorical(inp):
#inp = np.copy(inputData)
inp[inp == 255] = 1 # Line - comes first so that target data only has two categories
inp[inp == 65] = 2 # Box A
inp[inp == 150] = 3 # Box B
inp[inp == 215] = 4 # Dot
oldShape = inp.shape
inp = to_categorical(inp)
newShape = oldShape + (inp.shape[-1],)
return inp.reshape(newShape)
In [6]:
def imgSizeToBoxes(x):
return (x-3)/2
def lineFilterMatrixNP(imgWidth,imgHeight):
boxWidth = imgSizeToBoxes(imgWidth)
boxHeight = imgSizeToBoxes(imgHeight)
linesCnt = 2*boxWidth*boxHeight+boxWidth+boxHeight
mat = np.zeros((imgHeight, imgWidth), dtype=np.bool)
for idx in range(linesCnt):
y1 = idx / ((2*boxWidth) + 1)
if idx % ((2*boxWidth) + 1) < boxWidth:
# horizontal line
x1 = idx % ((2*boxWidth) + 1)
x2 = x1 + 1
y2 = y1
else:
# vertical line
x1 = idx % ((2*boxWidth) + 1) - boxWidth
x2 = x1
y2 = y1 + 1
px = x2 * 2 + y2 - y1
py = y2 * 2 + x2 - x1
mat[py,px] = 1
return mat
#lineFilterMatrixNP(13,11)
In [7]:
def loadDataset(datasetPath):
rawDataset = np.load(datasetPath)
x_train = rawDataset['x_train']
y_train = rawDataset['y_train']
x_train = dotsAndBoxesToCategorical(x_train)
y_train = y_train[:,lineFilterMatrixNP(y_train.shape[-1], y_train.shape[-2])]
y_train /= 255
return (x_train, y_train)
np.set_printoptions(precision=2)
In [8]:
(x_train, y_train) = loadDataset(datasetList[0])
print(x_train.shape)
print(y_train.shape)
In [9]:
kernelSize = (5,5)
filterCnt = 64
l2reg = 1e-4
resBlockCnt = 4
inputWidth = x_train.shape[-2]
inputHeight = x_train.shape[-3]
def build_residual_block(x, index):
in_x = x
res_name = "res"+str(index)
x = Conv2D(filters=filterCnt, kernel_size=kernelSize, padding="same",
data_format="channels_last", kernel_regularizer=l2(l2reg),
name=res_name+"_conv1_"+str(filterCnt))(x)
x = BatchNormalization(name=res_name+"_batchnorm1")(x)
x = Activation("relu",name=res_name+"_relu1")(x)
x = Conv2D(filters=filterCnt, kernel_size=kernelSize, padding="same",
data_format="channels_last", kernel_regularizer=l2(l2reg),
name=res_name+"_conv2-"+str(filterCnt))(x)
x = BatchNormalization(name="res"+str(index)+"_batchnorm2")(x)
x = Add(name=res_name+"_add")([in_x, x])
x = Activation("relu", name=res_name+"_relu2")(x)
return x
img_input = Input(shape=(None,None,5,))
x = Conv2D(filterCnt, kernelSize, padding='same', kernel_regularizer=l2(l2reg), name="input_conv")(img_input)
x = Activation("relu", name="input_relu")(x)
x = BatchNormalization()(x)
for i in range(resBlockCnt):
x = build_residual_block(x, i+1)
res_out = x
x = Conv2D(1, kernelSize, padding='same', kernel_regularizer=l2(l2reg), name="output_conv")(x)
x = LineFilterLayer(inputWidth, inputHeight)(x)
x = Activation("softmax", name="output_softmax")(x)
model = Model(inputs=img_input, outputs=x)
model.compile(optimizer='adam', loss='categorical_crossentropy')
model.summary()
model.save(modelPath)
In [10]:
#sess = K.get_session()
#sess = tf_debug.LocalCLIDebugWrapperSession(sess)
#K.set_session(sess)
iteration = 1
for datasetPath in datasetList:
print("cleaning up dataset")
del x_train
del y_train
gc.collect()
print("loading dataset " + datasetPath)
(x_train, y_train) = loadDataset(datasetPath)
print(x_train.shape)
print(y_train.shape)
# update the line filter layer to reflect new board size in dataset
LineFilterLayer.imgWidth = x_train.shape[-2]
LineFilterLayer.imgHeight = x_train.shape[-3]
model = load_model(modelPath, custom_objects={'LineFilterLayer':LineFilterLayer})
# Training
callbacks = []
checkpoint = ModelCheckpoint(filepath=modelPath+".checkpoint."+str(iteration), save_weights_only=False)
callbacks.append(checkpoint)
progbar = ProgbarLogger()
callbacks.append(progbar)
#tensorboard = TensorBoard(log_dir='model/log2', write_grads=True, write_graph=True, write_images=True, histogram_freq=1)
#callbacks.append(tensorboard)
model.fit(x_train, y_train, epochs=1, batch_size=64, callbacks=callbacks, validation_split=0.001)
model.save(modelPath)
iteration += 1
In [10]:
def linesToDotsAndBoxesImage(lines, imgWidth, imgHeight):
boxWidth = imgSizeToBoxes(imgWidth)
boxHeight = imgSizeToBoxes(imgHeight)
linesCnt = 2*boxWidth*boxHeight+boxWidth+boxHeight
mat = np.zeros((imgHeight, imgWidth), dtype=lines.dtype)
for idx in range(linesCnt):
y1 = idx / ((2*boxWidth) + 1)
if idx % ((2*boxWidth) + 1) < boxWidth:
# horizontal line
x1 = idx % ((2*boxWidth) + 1)
x2 = x1 + 1
y2 = y1
else:
# vertical line
x1 = idx % ((2*boxWidth) + 1) - boxWidth
x2 = x1
y2 = y1 + 1
px = x2 * 2 + y2 - y1
py = y2 * 2 + x2 - x1
mat[py,px] = lines[idx]
return mat
In [11]:
example = random.randrange(x_train.shape[0])
print("example: "+str(example))
input_data = x_train[example:example+1]
input_data_cat = x_train_cat[example:example+1]
prediction_lines = model.predict(input_data_cat)
prediction_lines_print = prediction_lines * 100
print(prediction_lines_print.astype(np.uint8))
print(np.sum(prediction_lines))
prediction = linesToDotsAndBoxesImage(prediction_lines[0], x_train.shape[2], x_train.shape[1])
# print input data
input_data_print = x_train[example,:,:]
input_data_print = input_data_print.astype(np.uint8)
print("input "+str(input_data_print.shape)+": ")
print(input_data_print)
# generate greyscale image data from input data
target_imgdata = x_train[example,:,:]
target_imgdata = target_imgdata.astype(np.uint8)
# print prediction
prediction_data_print = prediction * 100
prediction_data_print = prediction_data_print.astype(np.uint8)
print("prediction: ")
print(prediction_data_print)
# generate greyscale image data from prediction data
prediction_imgdata = prediction * 255
prediction_imgdata = prediction_imgdata.astype(np.uint8)
# merge image data in color channels
tmp = np.zeros((prediction.shape[0], prediction.shape[1]), dtype=np.uint8)
merged_imgdata = np.stack([target_imgdata, prediction_imgdata, tmp], axis=2)
#create image
img = Image.fromarray(merged_imgdata, 'RGB')
img = img.resize(size=(img.size[0]*10, img.size[1]*10))
img
Out[11]:
In [ ]: