def generate_sessions(self,
n_iters=float('inf'),
n_games=1,
reload_period=10):
"""
Generates sessions and records them to the database
:param n_iters: how many batches to generate (inf means generate forever)
:param n_games: how many games to maintain in parallel
:param reload_period: how often to read weights from database (every reload_period batches)
"""
pool = EnvPool(self.agent, self.make_env, n_games=n_games)
loop = count() if np.isinf(n_iters) else range(n_iters)
try:
self.db.load_all_params(self.agent)
except:
self.db.save_all_params(self.agent)
for epoch in loop:
if (epoch + 1) % reload_period == 0:
self.db.load_all_params(self.agent)
# play
prev_memory = list(pool.prev_memory_states)
observations, actions, rewards, memory, is_alive, info = pool.interact(
self.sequence_length)
# save sessions
for k in range(n_games):
self.db.record_session(observations[k], actions[k], rewards[k],
is_alive[k],
[mem[k] for mem in prev_memory])
def evaluate(self, n_games, *args, **kwargs):
"""
Play several full games and averages agent rewards. Prints some info unless verbose=False
:param n_games: how many games to play (successively without changing weights)
"""
self.db.load_all_params(self.agent, errors='warn')
return EnvPool(self.agent, self.make_env,
n_games=0).evaluate(n_games, *args, **kwargs)
qvalues_old = targetnet.output_layers
from agentnet.agent import Agent
#all together
agent = Agent(
observation_layers=observation_layer,
policy_estimators=(qvalues_layer, qvalues_old),
#agent_states={conv_rec:in_conv_rec},
action_layers=action_layer)
#Since it's a single lasagne network, one can get it's weights, output, etc
weights = lasagne.layers.get_all_params(qvalues_layer, trainable=True)
weights
from agentnet.experiments.openai_gym.pool import EnvPool
pool = EnvPool(agent, make_env, N_AGENTS)
## %%time
##interact for 7 ticks
#_,action_log,reward_log,_,_,_ = pool.interact(10)
#
#print('actions:')
#print(action_log[0])
#print("rewards")
#print(reward_log[0])
#load first sessions (this function calls interact and remembers sessions)
pool.update(SEQ_LENGTH)
#get agent's Qvalues obtained via experience replay
replay = pool.experience_replay
def test_space_invaders(
game_title='SpaceInvaders-v0',
n_parallel_games=3,
replay_seq_len=2,
):
"""
:param game_title: name of atari game in Gym
:param n_parallel_games: how many games we run in parallel
:param replay_seq_len: how long is one replay session from a batch
"""
atari = gym.make(game_title)
atari.reset()
# Game Parameters
n_actions = atari.action_space.n
observation_shape = (None, ) + atari.observation_space.shape
del atari
# ##### Agent observations
# image observation at current tick goes here
observation_layer = InputLayer(observation_shape, name="images input")
# reshape to [batch, color, x, y] to allow for convolutional layers to work correctly
observation_reshape = DimshuffleLayer(observation_layer, (0, 3, 1, 2))
# Agent memory states
window_size = 3
# prev state input
prev_window = InputLayer(
(None, window_size) + tuple(observation_reshape.output_shape[1:]),
name="previous window state")
# our window
window = WindowAugmentation(observation_reshape,
prev_window,
name="new window state")
memory_dict = {window: prev_window}
# ##### Neural network body
# you may use any other lasagne layers, including convolutions, batch_norms, maxout, etc
# pixel-wise maximum over the temporal window (to avoid flickering)
window_max = ExpressionLayer(window,
lambda a: a.max(axis=1),
output_shape=(None, ) +
window.output_shape[2:])
# a simple lasagne network (try replacing with any other lasagne network and see what works best)
nn = DenseLayer(window_max, num_units=50, name='dense0')
# Agent policy and action picking
q_eval = DenseLayer(nn,
num_units=n_actions,
nonlinearity=lasagne.nonlinearities.linear,
name="QEvaluator")
#fakes for a2c
policy_eval = DenseLayer(nn,
num_units=n_actions,
nonlinearity=lasagne.nonlinearities.softmax,
name="a2c action probas")
state_value_eval = DenseLayer(nn,
num_units=1,
nonlinearity=None,
name="a2c state values")
# resolver
resolver = ProbabilisticResolver(policy_eval, name="resolver")
# agent
agent = Agent(observation_layer, memory_dict,
(q_eval, policy_eval, state_value_eval), resolver)
# Since it's a single lasagne network, one can get it's weights, output, etc
weights = lasagne.layers.get_all_params(resolver, trainable=True)
# Agent step function
# # Create and manage a pool of atari sessions to play with
pool = EnvPool(agent, game_title, n_parallel_games)
observation_log, action_log, reward_log, _, _, _ = pool.interact(50)
# # experience replay pool
# Create an environment with all default parameters
env = SessionPoolEnvironment(observations=observation_layer,
actions=resolver,
agent_memories=agent.agent_states)
def update_pool(env, pool, n_steps=100):
""" a function that creates new sessions and ads them into the pool
throwing the old ones away entirely for simplicity"""
preceding_memory_states = list(pool.prev_memory_states)
# get interaction sessions
observation_tensor, action_tensor, reward_tensor, _, is_alive_tensor, _ = pool.interact(
n_steps=n_steps)
# load them into experience replay environment
env.load_sessions(observation_tensor, action_tensor, reward_tensor,
is_alive_tensor, preceding_memory_states)
# load first sessions
update_pool(env, pool, replay_seq_len)
# A more sophisticated way of training is to store a large pool of sessions and train on random batches of them.
# ### Training via experience replay
# get agent's Q-values, policy, etc obtained via experience replay
_env_states, _observations, _memories, _imagined_actions, estimators = agent.get_sessions(
env,
session_length=replay_seq_len,
batch_size=env.batch_size,
experience_replay=True,
)
(q_values_sequence, policy_sequence, value_sequence) = estimators
# Evaluating loss function
scaled_reward_seq = env.rewards
# For SpaceInvaders, however, not scaling rewards is at least working
elwise_mse_loss = 0.
#1-step algos
for algo in qlearning, sarsa:
elwise_mse_loss += algo.get_elementwise_objective(
q_values_sequence,
env.actions[0],
scaled_reward_seq,
env.is_alive,
gamma_or_gammas=0.99,
)
#qlearning_n_step
for n in (1, 3, replay_seq_len - 1, replay_seq_len, replay_seq_len + 1,
None):
elwise_mse_loss += qlearning.get_elementwise_objective(
q_values_sequence,
env.actions[0],
scaled_reward_seq,
env.is_alive,
gamma_or_gammas=0.99,
n_steps=n)
#a2c n_step
elwise_mse_loss += a2c.get_elementwise_objective(policy_sequence,
value_sequence[:, :, 0],
env.actions[0],
scaled_reward_seq,
env.is_alive,
gamma_or_gammas=0.99,
n_steps=3)
# compute mean over "alive" fragments
mse_loss = elwise_mse_loss.sum() / env.is_alive.sum()
# regularize network weights
reg_l2 = regularize_network_params(resolver, l2) * 10**-4
loss = mse_loss + reg_l2
# Compute weight updates
updates = lasagne.updates.adadelta(loss, weights, learning_rate=0.01)
# mean session reward
mean_session_reward = env.rewards.sum(axis=1).mean()
# # Compile train and evaluation functions
print('compiling')
train_fun = theano.function([], [loss, mean_session_reward],
updates=updates)
evaluation_fun = theano.function(
[], [loss, mse_loss, reg_l2, mean_session_reward])
print("I've compiled!")
# # Training loop
for epoch_counter in range(10):
update_pool(env, pool, replay_seq_len)
loss, avg_reward = train_fun()
full_loss, q_loss, l2_penalty, avg_reward_current = evaluation_fun()
print("epoch %i,loss %.5f, rewards: %.5f " %
(epoch_counter, full_loss, avg_reward_current))
print("rec %.3f reg %.3f" % (q_loss, l2_penalty))
def test_memory(
game_title='SpaceInvaders-v0',
n_parallel_games=3,
replay_seq_len=2,
):
"""
:param game_title: name of atari game in Gym
:param n_parallel_games: how many games we run in parallel
:param replay_seq_len: how long is one replay session from a batch
"""
atari = gym.make(game_title)
atari.reset()
# Game Parameters
n_actions = atari.action_space.n
observation_shape = (None, ) + atari.observation_space.shape
action_names = atari.get_action_meanings()
del atari
# ##### Agent observations
# image observation at current tick goes here
observation_layer = InputLayer(observation_shape, name="images input")
# reshape to [batch, color, x, y] to allow for convolutional layers to work correctly
observation_reshape = DimshuffleLayer(observation_layer, (0, 3, 1, 2))
# Agent memory states
memory_dict = OrderedDict([])
###Window
window_size = 3
# prev state input
prev_window = InputLayer(
(None, window_size) + tuple(observation_reshape.output_shape[1:]),
name="previous window state")
# our window
window = WindowAugmentation(observation_reshape,
prev_window,
name="new window state")
# pixel-wise maximum over the temporal window (to avoid flickering)
window_max = ExpressionLayer(window,
lambda a: a.max(axis=1),
output_shape=(None, ) +
window.output_shape[2:])
memory_dict[window] = prev_window
###Stack
#prev stack
stack_w, stack_h = 4, 5
stack_inputs = DenseLayer(observation_reshape, stack_w, name="prev_stack")
stack_controls = DenseLayer(observation_reshape,
3,
nonlinearity=lasagne.nonlinearities.softmax,
name="prev_stack")
prev_stack = InputLayer((None, stack_h, stack_w),
name="previous stack state")
stack = StackAugmentation(stack_inputs, prev_stack, stack_controls)
memory_dict[stack] = prev_stack
stack_top = lasagne.layers.SliceLayer(stack, 0, 1)
###RNN preset
prev_rnn = InputLayer((None, 16), name="previous RNN state")
new_rnn = RNNCell(prev_rnn, observation_reshape)
memory_dict[new_rnn] = prev_rnn
###GRU preset
prev_gru = InputLayer((None, 16), name="previous GRUcell state")
new_gru = GRUCell(prev_gru, observation_reshape)
memory_dict[new_gru] = prev_gru
###GRUmemorylayer
prev_gru1 = InputLayer((None, 15), name="previous GRUcell state")
new_gru1 = GRUMemoryLayer(15, observation_reshape, prev_gru1)
memory_dict[new_gru1] = prev_gru1
#LSTM with peepholes
prev_lstm0_cell = InputLayer(
(None, 13), name="previous LSTMCell hidden state [with peepholes]")
prev_lstm0_out = InputLayer(
(None, 13), name="previous LSTMCell output state [with peepholes]")
new_lstm0_cell, new_lstm0_out = LSTMCell(
prev_lstm0_cell,
prev_lstm0_out,
input_or_inputs=observation_reshape,
peepholes=True,
name="newLSTM1 [with peepholes]")
memory_dict[new_lstm0_cell] = prev_lstm0_cell
memory_dict[new_lstm0_out] = prev_lstm0_out
#LSTM without peepholes
prev_lstm1_cell = InputLayer(
(None, 14), name="previous LSTMCell hidden state [no peepholes]")
prev_lstm1_out = InputLayer(
(None, 14), name="previous LSTMCell output state [no peepholes]")
new_lstm1_cell, new_lstm1_out = LSTMCell(
prev_lstm1_cell,
prev_lstm1_out,
input_or_inputs=observation_reshape,
peepholes=False,
name="newLSTM1 [no peepholes]")
memory_dict[new_lstm1_cell] = prev_lstm1_cell
memory_dict[new_lstm1_out] = prev_lstm1_out
##concat everything
for i in [flatten(window_max), stack_top, new_rnn, new_gru, new_gru1]:
print(i.output_shape)
all_memory = concat([
flatten(window_max),
stack_top,
new_rnn,
new_gru,
new_gru1,
new_lstm0_out,
new_lstm1_out,
])
# ##### Neural network body
# you may use any other lasagne layers, including convolutions, batch_norms, maxout, etc
# a simple lasagne network (try replacing with any other lasagne network and see what works best)
nn = DenseLayer(all_memory, num_units=50, name='dense0')
# Agent policy and action picking
q_eval = DenseLayer(nn,
num_units=n_actions,
nonlinearity=lasagne.nonlinearities.linear,
name="QEvaluator")
# resolver
resolver = EpsilonGreedyResolver(q_eval, epsilon=0.1, name="resolver")
# agent
agent = Agent(observation_layer, memory_dict, q_eval, resolver)
# Since it's a single lasagne network, one can get it's weights, output, etc
weights = lasagne.layers.get_all_params(resolver, trainable=True)
# # Create and manage a pool of atari sessions to play with
pool = EnvPool(agent, game_title, n_parallel_games)
observation_log, action_log, reward_log, _, _, _ = pool.interact(50)
print(np.array(action_names)[np.array(action_log)[:3, :5]])
# # experience replay pool
# Create an environment with all default parameters
env = SessionPoolEnvironment(observations=observation_layer,
actions=resolver,
agent_memories=agent.agent_states)
def update_pool(env, pool, n_steps=100):
""" a function that creates new sessions and ads them into the pool
throwing the old ones away entirely for simplicity"""
preceding_memory_states = list(pool.prev_memory_states)
# get interaction sessions
observation_tensor, action_tensor, reward_tensor, _, is_alive_tensor, _ = pool.interact(
n_steps=n_steps)
# load them into experience replay environment
env.load_sessions(observation_tensor, action_tensor, reward_tensor,
is_alive_tensor, preceding_memory_states)
# load first sessions
update_pool(env, pool, replay_seq_len)
# A more sophisticated way of training is to store a large pool of sessions and train on random batches of them.
# ### Training via experience replay
# get agent's Q-values obtained via experience replay
_env_states, _observations, _memories, _imagined_actions, q_values_sequence = agent.get_sessions(
env,
session_length=replay_seq_len,
batch_size=env.batch_size,
experience_replay=True,
)
# Evaluating loss function
scaled_reward_seq = env.rewards
# For SpaceInvaders, however, not scaling rewards is at least working
elwise_mse_loss = qlearning.get_elementwise_objective(
q_values_sequence,
env.actions[0],
scaled_reward_seq,
env.is_alive,
gamma_or_gammas=0.99,
)
# compute mean over "alive" fragments
mse_loss = elwise_mse_loss.sum() / env.is_alive.sum()
# regularize network weights
reg_l2 = regularize_network_params(resolver, l2) * 10**-4
loss = mse_loss + reg_l2
# Compute weight updates
updates = lasagne.updates.adadelta(loss, weights, learning_rate=0.01)
# mean session reward
mean_session_reward = env.rewards.sum(axis=1).mean()
# # Compile train and evaluation functions
print('compiling')
train_fun = theano.function([], [loss, mean_session_reward],
updates=updates)
evaluation_fun = theano.function(
[], [loss, mse_loss, reg_l2, mean_session_reward])
print("I've compiled!")
# # Training loop
for epoch_counter in range(10):
update_pool(env, pool, replay_seq_len)
loss, avg_reward = train_fun()
full_loss, q_loss, l2_penalty, avg_reward_current = evaluation_fun()
print("epoch %i,loss %.5f, rewards: %.5f " %
(epoch_counter, full_loss, avg_reward_current))
print("rec %.3f reg %.3f" % (q_loss, l2_penalty))
def test_space_invaders(game_title='SpaceInvaders-v0',
n_parallel_games=3,
replay_seq_len=2,
):
"""
:param game_title: name of atari game in Gym
:param n_parallel_games: how many games we run in parallel
:param replay_seq_len: how long is one replay session from a batch
"""
atari = gym.make(game_title)
atari.reset()
# Game Parameters
n_actions = atari.action_space.n
observation_shape = (None,) + atari.observation_space.shape
del atari
# ##### Agent observations
# image observation at current tick goes here
observation_layer = InputLayer(observation_shape, name="images input")
# reshape to [batch, color, x, y] to allow for convolutional layers to work correctly
observation_reshape = DimshuffleLayer(observation_layer, (0, 3, 1, 2))
# Agent memory states
window_size = 3
# prev state input
prev_window = InputLayer((None, window_size) + tuple(observation_reshape.output_shape[1:]),
name="previous window state")
# our window
window = WindowAugmentation(observation_reshape,
prev_window,
name="new window state")
memory_dict = {window: prev_window}
# ##### Neural network body
# you may use any other lasagne layers, including convolutions, batch_norms, maxout, etc
# pixel-wise maximum over the temporal window (to avoid flickering)
window_max = ExpressionLayer(window,
lambda a: a.max(axis=1),
output_shape=(None,) + window.output_shape[2:])
# a simple lasagne network (try replacing with any other lasagne network and see what works best)
nn = DenseLayer(window_max, num_units=50, name='dense0')
# Agent policy and action picking
q_eval = DenseLayer(nn,
num_units=n_actions,
nonlinearity=lasagne.nonlinearities.linear,
name="QEvaluator")
#fakes for a2c
policy_eval = DenseLayer(nn,
num_units=n_actions,
nonlinearity=lasagne.nonlinearities.softmax,
name="a2c action probas")
state_value_eval = DenseLayer(nn,
num_units=1,
nonlinearity=None,
name="a2c state values")
# resolver
resolver = ProbabilisticResolver(policy_eval, name="resolver")
# agent
agent = Agent(observation_layer,
memory_dict,
(q_eval,policy_eval,state_value_eval), resolver)
# Since it's a single lasagne network, one can get it's weights, output, etc
weights = lasagne.layers.get_all_params(resolver, trainable=True)
# Agent step function
# # Create and manage a pool of atari sessions to play with
pool = EnvPool(agent,game_title, n_parallel_games)
observation_log, action_log, reward_log, _, _, _ = pool.interact(50)
# # experience replay pool
# Create an environment with all default parameters
env = SessionPoolEnvironment(observations=observation_layer,
actions=resolver,
agent_memories=agent.agent_states)
def update_pool(env, pool, n_steps=100):
""" a function that creates new sessions and ads them into the pool
throwing the old ones away entirely for simplicity"""
preceding_memory_states = list(pool.prev_memory_states)
# get interaction sessions
observation_tensor, action_tensor, reward_tensor, _, is_alive_tensor, _ = pool.interact(n_steps=n_steps)
# load them into experience replay environment
env.load_sessions(observation_tensor, action_tensor, reward_tensor, is_alive_tensor, preceding_memory_states)
# load first sessions
update_pool(env, pool, replay_seq_len)
# A more sophisticated way of training is to store a large pool of sessions and train on random batches of them.
# ### Training via experience replay
# get agent's Q-values, policy, etc obtained via experience replay
_env_states, _observations, _memories, _imagined_actions, estimators = agent.get_sessions(
env,
session_length=replay_seq_len,
batch_size=env.batch_size,
experience_replay=True,
)
(q_values_sequence,policy_sequence,value_sequence) = estimators
# Evaluating loss function
scaled_reward_seq = env.rewards
# For SpaceInvaders, however, not scaling rewards is at least working
elwise_mse_loss = 0.
#1-step algos
for algo in qlearning,sarsa:
elwise_mse_loss += algo.get_elementwise_objective(q_values_sequence,
env.actions[0],
scaled_reward_seq,
env.is_alive,
gamma_or_gammas=0.99, )
#qlearning_n_step
for n in (1,3,replay_seq_len-1, replay_seq_len, replay_seq_len+1,None):
elwise_mse_loss += qlearning.get_elementwise_objective(q_values_sequence,
env.actions[0],
scaled_reward_seq,
env.is_alive,
gamma_or_gammas=0.99,
n_steps=n)
#a2c n_step
elwise_mse_loss += a2c.get_elementwise_objective(policy_sequence,
value_sequence[:,:,0],
env.actions[0],
scaled_reward_seq,
env.is_alive,
gamma_or_gammas=0.99,
n_steps=3)
# compute mean over "alive" fragments
mse_loss = elwise_mse_loss.sum() / env.is_alive.sum()
# regularize network weights
reg_l2 = regularize_network_params(resolver, l2) * 10 ** -4
loss = mse_loss + reg_l2
# Compute weight updates
updates = lasagne.updates.adadelta(loss, weights, learning_rate=0.01)
# mean session reward
mean_session_reward = env.rewards.sum(axis=1).mean()
# # Compile train and evaluation functions
print('compiling')
train_fun = theano.function([], [loss, mean_session_reward], updates=updates)
evaluation_fun = theano.function([], [loss, mse_loss, reg_l2, mean_session_reward])
print("I've compiled!")
# # Training loop
for epoch_counter in range(10):
update_pool(env, pool, replay_seq_len)
loss, avg_reward = train_fun()
full_loss, q_loss, l2_penalty, avg_reward_current = evaluation_fun()
print("epoch %i,loss %.5f, rewards: %.5f " % (
epoch_counter, full_loss, avg_reward_current))
print("rec %.3f reg %.3f" % (q_loss, l2_penalty))
def test_memory(game_title='SpaceInvaders-v0',
n_parallel_games=3,
replay_seq_len=2,
):
"""
:param game_title: name of atari game in Gym
:param n_parallel_games: how many games we run in parallel
:param replay_seq_len: how long is one replay session from a batch
"""
atari = gym.make(game_title)
atari.reset()
# Game Parameters
n_actions = atari.action_space.n
observation_shape = (None,) + atari.observation_space.shape
del atari
# ##### Agent observations
# image observation at current tick goes here
observation_layer = InputLayer(observation_shape, name="images input")
# reshape to [batch, color, x, y] to allow for convolutional layers to work correctly
observation_reshape = DimshuffleLayer(observation_layer, (0, 3, 1, 2))
# Agent memory states
memory_dict = OrderedDict([])
###Window
window_size = 3
# prev state input
prev_window = InputLayer((None, window_size) + tuple(observation_reshape.output_shape[1:]),
name="previous window state")
# our window
window = WindowAugmentation(observation_reshape,
prev_window,
name="new window state")
# pixel-wise maximum over the temporal window (to avoid flickering)
window_max = ExpressionLayer(window,
lambda a: a.max(axis=1),
output_shape=(None,) + window.output_shape[2:])
memory_dict[window] = prev_window
###Stack
#prev stack
stack_w,stack_h = 4, 5
stack_inputs = DenseLayer(observation_reshape,stack_w,name="prev_stack")
stack_controls = DenseLayer(observation_reshape,3,
nonlinearity=lasagne.nonlinearities.softmax,
name="prev_stack")
prev_stack = InputLayer((None,stack_h,stack_w),
name="previous stack state")
stack = StackAugmentation(stack_inputs,prev_stack, stack_controls)
memory_dict[stack] = prev_stack
stack_top = lasagne.layers.SliceLayer(stack,0,1)
###RNN preset
prev_rnn = InputLayer((None,16),
name="previous RNN state")
new_rnn = RNNCell(prev_rnn,observation_reshape)
memory_dict[new_rnn] = prev_rnn
###GRU preset
prev_gru = InputLayer((None,16),
name="previous GRUcell state")
new_gru = GRUCell(prev_gru,observation_reshape)
memory_dict[new_gru] = prev_gru
###GRUmemorylayer
prev_gru1 = InputLayer((None,15),
name="previous GRUcell state")
new_gru1 = GRUMemoryLayer(15,observation_reshape,prev_gru1)
memory_dict[new_gru1] = prev_gru1
#LSTM with peepholes
prev_lstm0_cell = InputLayer((None,13),
name="previous LSTMCell hidden state [with peepholes]")
prev_lstm0_out = InputLayer((None,13),
name="previous LSTMCell output state [with peepholes]")
new_lstm0_cell,new_lstm0_out = LSTMCell(prev_lstm0_cell,prev_lstm0_out,
input_or_inputs = observation_reshape,
peepholes=True,name="newLSTM1 [with peepholes]")
memory_dict[new_lstm0_cell] = prev_lstm0_cell
memory_dict[new_lstm0_out] = prev_lstm0_out
#LSTM without peepholes
prev_lstm1_cell = InputLayer((None,14),
name="previous LSTMCell hidden state [no peepholes]")
prev_lstm1_out = InputLayer((None,14),
name="previous LSTMCell output state [no peepholes]")
new_lstm1_cell,new_lstm1_out = LSTMCell(prev_lstm1_cell,prev_lstm1_out,
input_or_inputs = observation_reshape,
peepholes=False,name="newLSTM1 [no peepholes]")
memory_dict[new_lstm1_cell] = prev_lstm1_cell
memory_dict[new_lstm1_out] = prev_lstm1_out
##concat everything
for i in [flatten(window_max),stack_top,new_rnn,new_gru,new_gru1]:
print(i.output_shape)
all_memory = concat([flatten(window_max),stack_top,new_rnn,new_gru,new_gru1,new_lstm0_out,new_lstm1_out,])
# ##### Neural network body
# you may use any other lasagne layers, including convolutions, batch_norms, maxout, etc
# a simple lasagne network (try replacing with any other lasagne network and see what works best)
nn = DenseLayer(all_memory, num_units=50, name='dense0')
# Agent policy and action picking
q_eval = DenseLayer(nn,
num_units=n_actions,
nonlinearity=lasagne.nonlinearities.linear,
name="QEvaluator")
# resolver
resolver = EpsilonGreedyResolver(q_eval, epsilon=0.1, name="resolver")
# agent
agent = Agent(observation_layer,
memory_dict,
q_eval, resolver)
# Since it's a single lasagne network, one can get it's weights, output, etc
weights = lasagne.layers.get_all_params(resolver, trainable=True)
# # Create and manage a pool of atari sessions to play with
pool = EnvPool(agent,game_title, n_parallel_games)
observation_log, action_log, reward_log, _, _, _ = pool.interact(50)
# # experience replay pool
# Create an environment with all default parameters
env = SessionPoolEnvironment(observations=observation_layer,
actions=resolver,
agent_memories=agent.agent_states)
def update_pool(env, pool, n_steps=100):
""" a function that creates new sessions and ads them into the pool
throwing the old ones away entirely for simplicity"""
preceding_memory_states = list(pool.prev_memory_states)
# get interaction sessions
observation_tensor, action_tensor, reward_tensor, _, is_alive_tensor, _ = pool.interact(n_steps=n_steps)
# load them into experience replay environment
env.load_sessions(observation_tensor, action_tensor, reward_tensor, is_alive_tensor, preceding_memory_states)
# load first sessions
update_pool(env, pool, replay_seq_len)
# A more sophisticated way of training is to store a large pool of sessions and train on random batches of them.
# ### Training via experience replay
# get agent's Q-values obtained via experience replay
_env_states, _observations, _memories, _imagined_actions, q_values_sequence = agent.get_sessions(
env,
session_length=replay_seq_len,
batch_size=env.batch_size,
experience_replay=True,
)
# Evaluating loss function
scaled_reward_seq = env.rewards
# For SpaceInvaders, however, not scaling rewards is at least working
elwise_mse_loss = qlearning.get_elementwise_objective(q_values_sequence,
env.actions[0],
scaled_reward_seq,
env.is_alive,
gamma_or_gammas=0.99, )
# compute mean over "alive" fragments
mse_loss = elwise_mse_loss.sum() / env.is_alive.sum()
# regularize network weights
reg_l2 = regularize_network_params(resolver, l2) * 10 ** -4
loss = mse_loss + reg_l2
# Compute weight updates
updates = lasagne.updates.adadelta(loss, weights, learning_rate=0.01)
# mean session reward
mean_session_reward = env.rewards.sum(axis=1).mean()
# # Compile train and evaluation functions
print('compiling')
train_fun = theano.function([], [loss, mean_session_reward], updates=updates)
evaluation_fun = theano.function([], [loss, mse_loss, reg_l2, mean_session_reward])
print("I've compiled!")
# # Training loop
for epoch_counter in range(10):
update_pool(env, pool, replay_seq_len)
loss, avg_reward = train_fun()
full_loss, q_loss, l2_penalty, avg_reward_current = evaluation_fun()
print("epoch %i,loss %.5f, rewards: %.5f " % (
epoch_counter, full_loss, avg_reward_current))
print("rec %.3f reg %.3f" % (q_loss, l2_penalty))