tensorlayer3/tensorlayer/rein.py

160 lines
4.7 KiB
Python

#! /usr/bin/python
# -*- coding: utf-8 -*-
import numpy as np
import tensorflow as tf
from six.moves import xrange
__all__ = [
'discount_episode_rewards',
'cross_entropy_reward_loss',
'log_weight',
'choice_action_by_probs',
]
def discount_episode_rewards(rewards=None, gamma=0.99, mode=0):
"""Take 1D float array of rewards and compute discounted rewards for an
episode. When encount a non-zero value, consider as the end a of an episode.
Parameters
----------
rewards : list
List of rewards
gamma : float
Discounted factor
mode : int
Mode for computing the discount rewards.
- If mode == 0, reset the discount process when encount a non-zero reward (Ping-pong game).
- If mode == 1, would not reset the discount process.
Returns
--------
list of float
The discounted rewards.
Examples
----------
>>> rewards = np.asarray([0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1])
>>> gamma = 0.9
>>> discount_rewards = tl.rein.discount_episode_rewards(rewards, gamma)
>>> print(discount_rewards)
[ 0.72899997 0.81 0.89999998 1. 0.72899997 0.81
0.89999998 1. 0.72899997 0.81 0.89999998 1. ]
>>> discount_rewards = tl.rein.discount_episode_rewards(rewards, gamma, mode=1)
>>> print(discount_rewards)
[ 1.52110755 1.69011939 1.87791049 2.08656716 1.20729685 1.34144104
1.49048996 1.65610003 0.72899997 0.81 0.89999998 1. ]
"""
if rewards is None:
raise Exception("rewards should be a list")
discounted_r = np.zeros_like(rewards, dtype=np.float32)
running_add = 0
for t in reversed(xrange(0, rewards.size)):
if mode == 0:
if rewards[t] != 0: running_add = 0
running_add = running_add * gamma + rewards[t]
discounted_r[t] = running_add
return discounted_r
def cross_entropy_reward_loss(logits, actions, rewards, name=None):
"""Calculate the loss for Policy Gradient Network.
Parameters
----------
logits : tensor
The network outputs without softmax. This function implements softmax inside.
actions : tensor or placeholder
The agent actions.
rewards : tensor or placeholder
The rewards.
Returns
--------
Tensor
The TensorFlow loss function.
Examples
----------
>>> states_batch_pl = tf.placeholder(tf.float32, shape=[None, D])
>>> network = InputLayer(states_batch_pl, name='input')
>>> network = DenseLayer(network, n_units=H, act=tf.ops.relu, name='relu1')
>>> network = DenseLayer(network, n_units=3, name='out')
>>> probs = network.outputs
>>> sampling_prob = tf.ops.softmax(probs)
>>> actions_batch_pl = tf.placeholder(tf.int32, shape=[None])
>>> discount_rewards_batch_pl = tf.placeholder(tf.float32, shape=[None])
>>> loss = tl.rein.cross_entropy_reward_loss(probs, actions_batch_pl, discount_rewards_batch_pl)
>>> train_op = tf.train.RMSPropOptimizer(learning_rate, decay_rate).minimize(loss)
"""
cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(labels=actions, logits=logits, name=name)
return tf.reduce_sum(tf.multiply(cross_entropy, rewards))
def log_weight(probs, weights, name='log_weight'):
"""Log weight.
Parameters
-----------
probs : tensor
If it is a network output, usually we should scale it to [0, 1] via softmax.
weights : tensor
The weights.
Returns
--------
Tensor
The Tensor after appling the log weighted expression.
"""
with tf.variable_scope(name):
exp_v = tf.reduce_mean(tf.log(probs) * weights)
return exp_v
def choice_action_by_probs(probs=(0.5, 0.5), action_list=None):
"""Choice and return an an action by given the action probability distribution.
Parameters
------------
probs : list of float.
The probability distribution of all actions.
action_list : None or a list of int or others
A list of action in integer, string or others. If None, returns an integer range between 0 and len(probs)-1.
Returns
--------
float int or str
The chosen action.
Examples
----------
>>> for _ in range(5):
>>> a = choice_action_by_probs([0.2, 0.4, 0.4])
>>> print(a)
0
1
1
2
1
>>> for _ in range(3):
>>> a = choice_action_by_probs([0.5, 0.5], ['a', 'b'])
>>> print(a)
a
b
b
"""
if action_list is None:
n_action = len(probs)
action_list = np.arange(n_action)
else:
if len(action_list) != len(probs):
raise Exception("number of actions should equal to number of probabilities.")
return np.random.choice(action_list, p=probs)