minigo/oneoffs/training_curve.py at cont · artasparks/minigo

Repository navigation

History

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

"""

Used to plot the accuracy of the policy and value networks in

predicting professional game moves and results over the course

of training. Check FLAGS for default values for what models to

load and what sgf files to parse.

Usage:

python training_curve.py

Sample 3 positions from each game

python training_curve.py --num_positions=3

Only grab games after 2005 (default is 2000)

python training_curve.py --min_year=2005

"""

import sys; sys.path.insert(0, '.')

import sgf_wrapper

import go

import dual_net

import preprocessing

import selfplay_mcts

import evaluation

import sgf_wrapper

import utils

import rl_loop

import os

import shipname

import main

import numpy as np

import itertools

import matplotlib.pyplot as plt

import coords

import pandas as pd

import tensorflow as tf

import sgf

import pdb

from tqdm import tqdm

from sgf_wrapper import sgf_prop

from gtp_wrapper import make_gtp_instance, MCTSPlayer

from utils import logged_timer as timer

from tensorflow import gfile

tf.app.flags.DEFINE_string(

"sgf_dir", "sgf/baduk_db/", "sgf database")

tf.app.flags.DEFINE_string("model_dir", "saved_models",

"Where the model files are saved")

tf.app.flags.DEFINE_string("plot_dir", "data", "Where to save the plots.")

tf.app.flags.DEFINE_integer("min_year", "2000",

"Only take sgf games with date >= min_year")

tf.app.flags.DEFINE_string("komi", "7.5",

"Only take sgf games with given komi")

tf.app.flags.DEFINE_integer("idx_start", 150,

"Only take models after given idx")

tf.app.flags.DEFINE_integer("num_positions", 1,

"How many positions from each game to sample from.")

tf.app.flags.DEFINE_integer("eval_every", 5,

"Eval every k models to generate the curve")

FLAGS = tf.app.flags.FLAGS

def get_model_paths(model_dir):

'''Returns all model paths in the model_dir.'''

all_models = gfile.Glob(os.path.join(model_dir, '*.meta'))

model_filenames = [os.path.basename(m) for m in all_models]

model_numbers_names = [

(shipname.detect_model_num(m), shipname.detect_model_name(m))

for m in model_filenames]

model_names = sorted(model_numbers_names)

return [os.path.join(model_dir, name[1]) for name in model_names]

def load_player(model_path):

print("Loading weights from %s ... " % model_path)

with timer("Loading weights from %s ... " % model_path):

network = dual_net.DualNetwork(model_path)

network.name = os.path.basename(model_path)

player = MCTSPlayer(network, verbosity=2)

return player

def restore_params(model_path, player):

with player.network.sess.graph.as_default():

player.network.initialize_weights(model_path)

def batch_run_many(player, positions, batch_size=100):

"""Used to avoid a memory oveflow issue when running the network

on too many positions. TODO: This should be a member function of

player.network?"""

prob_list = []

value_list = []

for idx in range(0, len(positions), batch_size):

probs, values = player.network.run_many(positions[idx:idx+batch_size])

prob_list.append(probs)

value_list.append(values)

return np.concatenate(prob_list, axis=0), np.concatenate(value_list, axis=0)

def eval_player(player, positions, moves, results):

probs, values = batch_run_many(player, positions)

policy_moves = [coords.from_flat(c) for c in np.argmax(probs, axis=1)]

top_move_agree = [moves[idx] == policy_moves[idx] for idx in range(len(moves))]

square_err = (values - results)**2/4

return top_move_agree, square_err

def get_sgf_props(sgf_path):

with open(sgf_path) as f:

sgf_contents = f.read()

collection = sgf.parse(sgf_contents)

game = collection.children[0]

props = game.root.properties

return props

def parse_sgf(sgf_path):

with open(sgf_path) as f:

sgf_contents = f.read()

collection = sgf.parse(sgf_contents)

game = collection.children[0]

props = game.root.properties

assert int(sgf_prop(props.get('GM', ['1']))) == 1, "Not a Go SGF!"

result = utils.parse_game_result(sgf_prop(props.get('RE')))

positions, moves = zip(*[(p.position, p.next_move) for p in sgf_wrapper.replay_sgf(sgf_contents)])

return positions, moves, result, props

def check_year(props, year):

if year is None:

return True

if props.get('DT') is None:

return False

try:

#Most sgf files in this database have dates of the form

#"2005-01-15", but there are some rare exceptions like

#"Broadcasted on 2005-01-15.

year_sgf = int(props.get('DT')[0][:4])

except:

return False

#TODO: better to use datetime comparison here?

return year_sgf >= year

def check_komi(props, komi_str):

if komi_str is None:

return True

if props.get('KM') is None:

return False

return props.get('KM')[0] == komi_str

def find_and_filter_sgf_files(base_dir, min_year = None, komi = None):

base_dir = "sgf/baduk_db"

sgf_files = []

count = 0

print("Finding all sgf files in {} with year >= {} and komi = {}".format(base_dir, min_year, komi))

for i, (dirpath, dirnames, filenames) in tqdm(enumerate(os.walk(base_dir))):

for filename in filenames:

count+=1

if count%5000 == 0:

print("Parsed {}, Found {}".format(count, len(sgf_files)))

if filename.endswith('.sgf'):

path = os.sep.join([dirpath, filename])

props = get_sgf_props(path)

if check_year(props, min_year) and check_komi(props, komi):

sgf_files.append(path)

print("Found {} sgf files matching filters".format(len(sgf_files)))

return sgf_files

def sample_positions_from_games(sgf_files, num_positions=1):

pos_data = []

move_data = []

result_data = []

move_idxs = []

fail_count = 0

for i, path in tqdm(enumerate(sgf_files)):

if i % 1000 == 0:

print(i)

try:

positions, moves, result, props = parse_sgf(path)

except:

fail_count+=1

print("Fail count: {}".format(fail_count))

continue

#add entire game

if num_positions== -1:

pos_data.extend(positions)

move_data.extend(moves)

move_idxs.extend(range(len(positions)))

result_data.extend([result for i in range(len(positions))])

else:

for idx in np.random.choice(len(positions), num_positions):

pos_data.append(positions[idx])

move_data.append(moves[idx])

result_data.append(result)

move_idxs.append(idx)

print("Sampled {} positions, failed to parse {} files".format(len(pos_data), fail_count))

return pos_data, move_data, result_data, move_idxs

def get_training_curve_data(model_dir, pos_data, move_data, result_data, idx_start=150, eval_every=10):

model_paths = get_model_paths(model_dir)

df = pd.DataFrame()

player=None

print("Evaluating models {}-{}, eval_every={}".format(idx_start, len(model_paths), eval_every))

for idx in tqdm(range(idx_start, len(model_paths), eval_every)):

if player:

restore_params(model_paths[idx], player)

else:

player = load_player(model_paths[idx])

correct, squared_errors = eval_player(player=player, positions=pos_data, moves=move_data, results=result_data)

avg_acc = np.mean(correct)

avg_mse = np.mean(squared_errors)

print("Model: {}, acc: {}, mse: {}".format(model_paths[idx], avg_acc, avg_mse))

df = df.append({"num": idx, "acc": avg_acc, "mse": avg_mse}, ignore_index=True)

return df

def save_plots(data_dir, df):

plt.plot(df["num"], df["acc"])

plt.xlabel("Model idx")

plt.ylabel("Accuracy")

plt.title("Accuracy in Predicting Professional Moves")

plot_path = os.sep.join([data_dir, "move_acc.pdf"])

plt.savefig(plot_path)

plt.figure()

plt.plot(df["num"], df["mse"])

plt.xlabel("Model idx")

plt.ylabel("MSE/4")

plt.title("MSE in predicting outcome")

plot_path = os.sep.join([data_dir, "value_mse.pdf"])

plt.savefig(plot_path)

def main(unusedargv):

sgf_files = find_and_filter_sgf_files(FLAGS.sgf_dir, FLAGS.min_year, FLAGS.komi)

pos_data, move_data, result_data, move_idxs = sample_positions_from_games(sgf_files=sgf_files, num_positions=FLAGS.num_positions)

df = get_training_curve_data(FLAGS.model_dir, pos_data, move_data, result_data, FLAGS.idx_start, eval_every=FLAGS.eval_every)

save_plots(FLAGS.plot_dir, df)

FLAGS = tf.app.flags.FLAGS

if __name__ == "__main__":

tf.app.run(main)

FilesExpand file tree

training_curve.py

Latest commit

History

training_curve.py

File metadata and controls