def run(args):
# create dummy environment to be able to create model
env = gym.make(args.environment)
assert isinstance(env.observation_space, Box)
assert isinstance(env.action_space, Discrete)
print("Observation space:", env.observation_space)
print("Action space:", env.action_space)
# create main model
model = create_model(env, args)
model.summary()
env.close()
# force runner processes to use cpu
os.environ["CUDA_VISIBLE_DEVICES"] = ""
# for better compatibility with Theano and Tensorflow
multiprocessing.set_start_method('spawn')
# create shared buffer for sharing weights
blob = pickle.dumps(model.get_weights(), pickle.HIGHEST_PROTOCOL)
shared_buffer = Array('c', len(blob))
shared_buffer.raw = blob
# create fifos and threads for all runners
fifos = []
for i in range(args.num_runners):
fifo = Queue(args.queue_length)
fifos.append(fifo)
process = Process(target=runner, args=(shared_buffer, fifo, args))
process.start()
# start trainer in main thread
trainer(model, fifos, shared_buffer, args)
def run(args):
# create dummy environment to be able to create model
env = gym.make(args.environment)
assert isinstance(env.observation_space, Box)
assert isinstance(env.action_space, Discrete)
print("Observation space:", env.observation_space)
print("Action space:", env.action_space)
# create main model
model = create_model(env, args)
model.summary()
env.close()
# for better compatibility with Theano and Tensorflow
multiprocessing.set_start_method('spawn')
# create shared buffer for sharing weights
blob = pickle.dumps(model.get_weights(), pickle.HIGHEST_PROTOCOL)
shared_buffer = Array('c', len(blob))
shared_buffer.raw = blob
# force runner processes to use cpu
os.environ["CUDA_VISIBLE_DEVICES"] = ""
# create fifos and threads for all runners
fifos = []
for i in range(args.num_runners):
fifo = Queue(args.queue_length)
fifos.append(fifo)
process = Process(target=runner, args=(shared_buffer, fifo, args))
process.start()
# start trainer in main thread
trainer(model, fifos, shared_buffer, args)
def run(args):
# create dummy environment to be able to create model
env = create_env(args.env_id)
assert isinstance(env.observation_space, Box)
assert isinstance(env.action_space, Discrete)
print("Observation space: " + str(env.observation_space))
print("Action space: " + str(env.action_space))
# create main model
model = create_model(env,
batch_size=args.num_runners,
num_steps=args.num_local_steps)
model.summary()
env.close()
# for better compatibility with Theano and Tensorflow
multiprocessing.set_start_method('spawn')
# create shared buffer for sharing weights
blob = pickle.dumps(model.get_weights(), pickle.HIGHEST_PROTOCOL)
shared_buffer = Array('c', len(blob))
shared_buffer.raw = blob
# force runner processes to use cpu, child processes inherit environment variables
os.environ["CUDA_VISIBLE_DEVICES"] = ""
# create fifos and processes for all runners
fifos = []
for i in range(args.num_runners):
fifo = Queue(args.queue_length)
fifos.append(fifo)
process = Process(target=runner,
args=(shared_buffer, fifo,
args.num_timesteps // args.num_runners,
args.monitor and i == 0, args))
process.start()
# start trainer in main thread
trainer(model, fifos, shared_buffer, args)
print("All done")
def main():
global WORDLIST, HASHFILE, words, result, curr, total, num_words, curr_words
#
# process files
#
print("[*] reading hashes...")
hashes = open(HASHFILE, 'r')
hashlist = []
for line in hashes:
data = line.split(":")
if len(data) > 1:
hashv = data[0].strip()
salt = data[1].strip()
hashlist.append((hashv, salt))
hashes.close()
print("[*] parsing wordlist...")
words = Array('c', SHARED_MEM_SIZE, lock=False) # allocate shared memory segment
# get line count
wordlist_file = open(WORDLIST, 'r')
lines = 0
for line in wordlist_file:
lines += 1
total = lines * len(hashlist)
curr = Value('i', 0)
curr_words = Value('i', 0)
wordlist_file.seek(0) # get back to beginning
#
# crack
#
print("[*] beginning cracking")
pool = Pool(processes=NUM_PROCESSES)
results = []
current_char_count = 0
words_raw = ""
for line in wordlist_file:
length = len(line)
if length + current_char_count < SHARED_MEM_SIZE:
words_raw += line
current_char_count += length
else:
print("[*] next round")
curr_words.value = len(words_raw.split("\n"))
words.raw = words_raw + (SHARED_MEM_SIZE - len(words_raw)) * '0' # clear space
words_raw = line
current_char_count = length
# let workers do work!
results.extend(pool.map(entry, hashlist))
# remove cracked hashes
# TODO
print("[*] final round")
curr_words.value = len(words_raw.split("\n"))
words.raw = words_raw + (SHARED_MEM_SIZE - len(words_raw)) * '0'
results.extend(pool.map(entry, hashlist))
print("[*] done")
for result in results:
if result is not None:
print("%s:%s" % (result))
cv2.destroyAllWindows()
done.value = 1
if __name__ == '__main__':
#from tensorflow.python.client import device_lib
#print(device_lib.list_local_devices())
# Command-line argument parsing.
parser = setup_parser()
args = parser.parse_args()
# set up interprocess communication buffers
img = 128 * np.ones((args.frame_height, args.frame_width, 3), dtype=np.uint8)
buf = img.tostring()
last_frame = Array('c', len(buf))
last_frame.raw = buf
done = Value('i', 0)
# launch capture process
c = Process(name='capture', target=capture , args=(last_frame, done, args))
c.start()
# launch processing process
p = Process(name='processing', target=processing, args=(last_frame, done, args))
p.start()
# wait for processes to finish
p.join()
c.join()
