gaspersic/baselines
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

fix 'async' is a reserved word in Python >= 3.7 (#495) (#542)

Browse Source
This commit is contained in:
Tom
2018-08-28 03:36:43 +08:00
committed by pzhokhov
parent e92a6ad8f4
commit c6c0f45cb1
4 changed files with 7 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
baselines/acktr/acktr_cont.py
View File

@@ -54,7 +54,7 @@ def learn(env, policy, vf, gamma, lam, timesteps_per_batch, num_timesteps,
stepsize = tf.Variable(initial_value=np.float32(np.array(0.03)), name='stepsize') stepsize = tf.Variable(initial_value=np.float32(np.array(0.03)), name='stepsize')
inputs, loss, loss_sampled = policy.update_info inputs, loss, loss_sampled = policy.update_info
optim = kfac.KfacOptimizer(learning_rate=stepsize, cold_lr=stepsize*(1-0.9), momentum=0.9, kfac_update=2,\ optim = kfac.KfacOptimizer(learning_rate=stepsize, cold_lr=stepsize*(1-0.9), momentum=0.9, kfac_update=2,\
epsilon=1e-2, stats_decay=0.99, async=1, cold_iter=1, epsilon=1e-2, stats_decay=0.99, async_=1, cold_iter=1,
weight_decay_dict=policy.wd_dict, max_grad_norm=None) weight_decay_dict=policy.wd_dict, max_grad_norm=None)
pi_var_list = [] pi_var_list = []
for var in tf.trainable_variables(): for var in tf.trainable_variables():

6
baselines/acktr/acktr_disc.py
View File

@@ -58,7 +58,7 @@ class Model(object):
with tf.device('/gpu:0'): with tf.device('/gpu:0'):
self.optim = optim = kfac.KfacOptimizer(learning_rate=PG_LR, clip_kl=kfac_clip,\ self.optim = optim = kfac.KfacOptimizer(learning_rate=PG_LR, clip_kl=kfac_clip,\
momentum=0.9, kfac_update=1, epsilon=0.01,\ momentum=0.9, kfac_update=1, epsilon=0.01,\
stats_decay=0.99, async=1, cold_iter=10, max_grad_norm=max_grad_norm) stats_decay=0.99, async_=1, cold_iter=10, max_grad_norm=max_grad_norm)
update_stats_op = optim.compute_and_apply_stats(joint_fisher_loss, var_list=params) update_stats_op = optim.compute_and_apply_stats(joint_fisher_loss, var_list=params)
train_op, q_runner = optim.apply_gradients(list(zip(grads,params))) train_op, q_runner = optim.apply_gradients(list(zip(grads,params)))
@@ -97,7 +97,7 @@ def learn(network, env, seed, total_timesteps=int(40e6), gamma=0.99, log_interva
kfac_clip=0.001, save_interval=None, lrschedule='linear', load_path=None, **network_kwargs): kfac_clip=0.001, save_interval=None, lrschedule='linear', load_path=None, **network_kwargs):
set_global_seeds(seed) set_global_seeds(seed)
if network == 'cnn': if network == 'cnn':
network_kwargs['one_dim_bias'] = True network_kwargs['one_dim_bias'] = True
@@ -115,7 +115,7 @@ def learn(network, env, seed, total_timesteps=int(40e6), gamma=0.99, log_interva
with open(osp.join(logger.get_dir(), 'make_model.pkl'), 'wb') as fh: with open(osp.join(logger.get_dir(), 'make_model.pkl'), 'wb') as fh:
fh.write(cloudpickle.dumps(make_model)) fh.write(cloudpickle.dumps(make_model))
model = make_model() model = make_model()
if load_path is not None: if load_path is not None:
model.load(load_path) model.load(load_path)

4
baselines/acktr/kfac.py
View File

@@ -10,14 +10,14 @@ KFAC_DEBUG = False
class KfacOptimizer(): class KfacOptimizer():
def __init__(self, learning_rate=0.01, momentum=0.9, clip_kl=0.01, kfac_update=2, stats_accum_iter=60, full_stats_init=False, cold_iter=100, cold_lr=None, async=False, async_stats=False, epsilon=1e-2, stats_decay=0.95, blockdiag_bias=False, channel_fac=False, factored_damping=False, approxT2=False, use_float64=False, weight_decay_dict={},max_grad_norm=0.5): def __init__(self, learning_rate=0.01, momentum=0.9, clip_kl=0.01, kfac_update=2, stats_accum_iter=60, full_stats_init=False, cold_iter=100, cold_lr=None, async_=False, async_stats=False, epsilon=1e-2, stats_decay=0.95, blockdiag_bias=False, channel_fac=False, factored_damping=False, approxT2=False, use_float64=False, weight_decay_dict={},max_grad_norm=0.5):
self.max_grad_norm = max_grad_norm self.max_grad_norm = max_grad_norm
self._lr = learning_rate self._lr = learning_rate
self._momentum = momentum self._momentum = momentum
self._clip_kl = clip_kl self._clip_kl = clip_kl
self._channel_fac = channel_fac self._channel_fac = channel_fac
self._kfac_update = kfac_update self._kfac_update = kfac_update
self._async = async self._async = async_
self._async_stats = async_stats self._async_stats = async_stats
self._epsilon = epsilon self._epsilon = epsilon
self._stats_decay = stats_decay self._stats_decay = stats_decay

2
baselines/acktr/value_functions.py
View File

@@ -21,7 +21,7 @@ class NeuralNetValueFunction(object):
self._predict = U.function([X], vpred_n) self._predict = U.function([X], vpred_n)
optim = kfac.KfacOptimizer(learning_rate=0.001, cold_lr=0.001*(1-0.9), momentum=0.9, \ optim = kfac.KfacOptimizer(learning_rate=0.001, cold_lr=0.001*(1-0.9), momentum=0.9, \
clip_kl=0.3, epsilon=0.1, stats_decay=0.95, \ clip_kl=0.3, epsilon=0.1, stats_decay=0.95, \
async=1, kfac_update=2, cold_iter=50, \ async_=1, kfac_update=2, cold_iter=50, \
weight_decay_dict=wd_dict, max_grad_norm=None) weight_decay_dict=wd_dict, max_grad_norm=None)
vf_var_list = [] vf_var_list = []
for var in tf.trainable_variables(): for var in tf.trainable_variables():