I need to make a cache environment for my DQN Agent. I need to do network caching, when a file is needed it goes on cache if there is space. If the file is in cache, the agent has a reward. If the file is not in cache and the agent doesn't put it in cache, it has a bad reward.
My environment cache doesn't work, what is the problem? someone can make it work ?
import numpy as np
import time
import sys
from gym.spaces import Discrete, Box
import pandas as pd
class CacheProva (object):
def __init__(self, cache_max):
# Counters
self.total_count = 0
self.miss_count = 0
self.action_space = [0, 1, 2]
self.df = pd.read_csv('data/size.csv')
self.df_cache = pd.DataFrame(columns=['movie', 'dim'])
self.observation = 0 #sarebbe self.cahe_size
self.n_actions = len(self.action_space)#self.observation + 1
self.n_features = 0
def miss_rate(self):
return self.miss_count / self.total_count
def step(self, action):
if self.hasDone():
raise ValueError("Simulation has finished, use reset() to restart simulation.")
for index, row in self.df.iterrows():
self.total_count += 1
for i, r in self.df_cache.iterrows():
#self.total_count += 1
if action == 0:#gia presente nella cache
self.df.at[index, row] == self.df_cache.at[i, r]
reward = 1
if action == 1:#mettere nella cache
reward = 0
self.df_cache.append(self.df.at[index, row])#aggiungere a df_cache
self.observation += self.df.at['size']
elif action == 2:#non è presente e non si mette
reward = -1
self.miss_count += 1
return self.observation, reward, self.total_count, self.miss_count
#self.cureindex
#if id è in cache size o cache size è vuota, action 0 con reward 1
#if id non è in cache size e non c'è posto, action 1 la sostituisci reward 0
# if id non è in cache size e non c'è posto, action 2 non fai nulla reward -1
def render(self):
pass
def reset(self):
self.total_count = 0
self.miss_count = 0
self.observation = 0
return self.total_count, self.miss_count, self.observation
#return self.df_cache
def hasDone(self):
#return self.cur_index == self.df.shape[0]
pass
def display(self):
print(self.observation)
The csv file is a chronology of netflix users, of the film and series watched.
this is my DQN :
import numpy as np
import tensorflow.compat.v1 as tf
tf.disable_v2_behavior()
np.random.seed(1)
tf.random.set_random_seed(1)
# Deep Q Network off-policy
class DeepQNetwork:
def __init__(
self,
n_actions,
n_features,
learning_rate=0.01,
reward_decay=0.9,
e_greedy=0.9,
replace_target_iter=300,
memory_size=500,
batch_size=32,
e_greedy_increment=None,
output_graph=False,
):
self.n_actions = n_actions
self.n_features = n_features
self.lr = learning_rate
self.gamma = reward_decay
self.epsilon_max = e_greedy
self.replace_target_iter = replace_target_iter
self.memory_size = memory_size
self.batch_size = batch_size
self.epsilon_increment = e_greedy_increment
self.epsilon = 0 if e_greedy_increment is not None else self.epsilon_max
# total learning step
self.learn_step_counter = 0
# initialize zero memory [s, a, r, s_]
self.memory = np.zeros((self.memory_size, n_features * 2 + 2))
# consist of [target_net, evaluate_net]
self._build_net()
t_params = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope='target_net')
e_params = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope='eval_net')
with tf.variable_scope('hard_replacement'):
self.target_replace_op = [tf.assign(t, e) for t, e in zip(t_params, e_params)]
self.sess = tf.Session()
if output_graph:
# $ tensorboard --logdir=logs
tf.summary.FileWriter("logs/", self.sess.graph)
self.sess.run(tf.global_variables_initializer())
self.cost_his = []
def _build_net(self):
# ------------------ all inputs ------------------------
self.s = tf.placeholder(tf.float32, [None, self.n_features], name='s') # input State
self.s_ = tf.placeholder(tf.float32, [None, self.n_features], name='s_') # input Next State
self.r = tf.placeholder(tf.float32, [None, ], name='r') # input Reward
self.a = tf.placeholder(tf.int32, [None, ], name='a') # input Action
w_initializer, b_initializer = tf.random_normal_initializer(0., 0.3), tf.constant_initializer(0.1)
# ------------------ build evaluate_net ------------------
with tf.variable_scope('eval_net'):
e1 = tf.layers.dense(self.s, 20, tf.nn.relu, kernel_initializer=w_initializer,
bias_initializer=b_initializer, name='e1')
self.q_eval = tf.layers.dense(e1, self.n_actions, kernel_initializer=w_initializer,
bias_initializer=b_initializer, name='q')
# ------------------ build target_net ------------------
with tf.variable_scope('target_net'):
t1 = tf.layers.dense(self.s_, 20, tf.nn.relu, kernel_initializer=w_initializer,
bias_initializer=b_initializer, name='t1')
self.q_next = tf.layers.dense(t1, self.n_actions, kernel_initializer=w_initializer,
bias_initializer=b_initializer, name='t2')
with tf.variable_scope('q_target'):
q_target = self.r + self.gamma * tf.reduce_max(self.q_next, axis=1, name='Qmax_s_') # shape=(None, )
self.q_target = tf.stop_gradient(q_target)
with tf.variable_scope('q_eval'):
a_indices = tf.stack([tf.range(tf.shape(self.a)[0], dtype=tf.int32), self.a], axis=1)
self.q_eval_wrt_a = tf.gather_nd(params=self.q_eval, indices=a_indices) # shape=(None, )
with tf.variable_scope('loss'):
self.loss = tf.reduce_mean(tf.squared_difference(self.q_target, self.q_eval_wrt_a, name='TD_error'))
with tf.variable_scope('train'):
self._train_op = tf.train.RMSPropOptimizer(self.lr).minimize(self.loss)
def store_transition(self, s, a, r, s_):
if not hasattr(self, 'memory_counter'):
self.memory_counter = 0
transition = np.hstack((s, [a, r], s_))
# replace the old memory with new memory
index = self.memory_counter % self.memory_size
self.memory[index, :] = transition
self.memory_counter += 1
def choose_action(self, observation):
# to have batch dimension when feed into tf placeholder
observation = observation[np.newaxis, :]
if np.random.uniform() < self.epsilon:
# forward feed the observation and get q value for every actions
actions_value = self.sess.run(self.q_eval, feed_dict={self.s: observation})
action = np.argmax(actions_value)
else:
action = np.random.randint(0, self.n_actions)
return action
def learn(self):
# check to replace target parameters
if self.learn_step_counter % self.replace_target_iter == 0:
self.sess.run(self.target_replace_op)
print('\ntarget_params_replaced\n')
# sample batch memory from all memory
if self.memory_counter > self.memory_size:
sample_index = np.random.choice(self.memory_size, size=self.batch_size)
else:
sample_index = np.random.choice(self.memory_counter, size=self.batch_size)
batch_memory = self.memory[sample_index, :]
_, cost = self.sess.run(
[self._train_op, self.loss],
feed_dict={
self.s: batch_memory[:, :self.n_features],
self.a: batch_memory[:, self.n_features],
self.r: batch_memory[:, self.n_features + 1],
self.s_: batch_memory[:, -self.n_features:],
})
self.cost_his.append(cost)
# increasing epsilon
self.epsilon = self.epsilon + self.epsilon_increment if self.epsilon < self.epsilon_max else self.epsilon_max
self.learn_step_counter += 1
def plot_cost(self):
import matplotlib.pyplot as plt
plt.plot(np.arange(len(self.cost_his)), self.cost_his)
plt.ylabel('Cost')
plt.xlabel('training steps')
plt.show()
if __name__ == '__main__':
DQN = DeepQNetwork(3,4, output_graph=False)
Related
Hello my DDPG model that I have implemented in TF 2 get's horrible results at every env on openai-gym that has continuous actions I need help to find what's the problem. I run this on my GPU. On env Pendulum I get -1200/-1000 rewards on every episode. This code is from a course I took on udemy but it was written in TF1.x and I rewrote it in TF2 but his TF1.x implementation had better results. Here is the code:
import tensorflow as tf
import numpy as np
import os
import gym
import random
import matplotlib.pyplot as plt
from tensorflow.keras.layers import Input, Dense, concatenate
from tensorflow.keras.models import Model
class ReplayBuffer():
def __init__(self, obs_dim, act_dim, size):
self.obs1_buf = np.zeros([size, obs_dim, ], dtype=np.float32)
self.obs2_buf = np.zeros([size, obs_dim, ], dtype=np.float32)
self.act_buf = np.zeros([size, act_dim], dtype=np.float32)
self.reward_buf = np.zeros(size, dtype=np.float32)
self.done_buf = np.zeros(size, dtype=np.float32)
self.current = 0
self.count = 0
self.size = size
def add_experience(self, state, action, reward, next_state, done):
self.obs1_buf[self.current] = state
self.act_buf[self.current] = action
self.reward_buf[self.current] = reward
self.obs2_buf[self.current] = next_state
self.done_buf[self.current] = done
self.current = (self.current + 1) % self.size
self.count = min(self.count+1, self.size)
def sample_batch(self, batch_size=32):
idx = np.random.randint(0, self.count, size=batch_size)
return dict(s=self.obs1_buf[idx],
s2=self.obs2_buf[idx],
a=self.act_buf[idx],
r=self.reward_buf[idx],
d=self.done_buf[idx])
class DDPG():
def __init__(self, env, num_states, num_actions, action_max):
self.env = env
self.num_states = num_states
self.num_actions = num_actions
self.action_max = action_max
self.gamma = 0.99
self.decay = 0.995
self.mu_optimizer = tf.keras.optimizers.Adam(learning_rate=0.01)
self.q_optimizer = tf.keras.optimizers.Adam(learning_rate=0.01)
def mu_model(hidden_layers):
inp = Input(shape=(self.num_states, ))
x = inp
for layers in hidden_layers[:-1]:
x = Dense(layers, activation='relu')(x)
x = Dense(hidden_layers[-1], activation='tanh')(x)
mu_model = Model(inp, x)
return mu_model
self.mu_model = mu_model([300, self.num_actions])
def q_model(inp_state, inp_act, hidden_layers):
inp_state = Input(shape=(inp_state, ))
inp_mu = Input(shape=(inp_act, ))
inp = concatenate([inp_state, inp_mu])
x = inp
for layers in hidden_layers[:-1]:
x = Dense(layers, activation='relu')(x)
x = Dense(hidden_layers[-1], activation='linear')(x)
q_model = Model([inp_state, inp_mu], x)
return q_model
self.q_model = q_model(self.num_states, self.num_actions, hidden_layers=[300, 1])
self.q_target_model = q_model(self.num_states, self.num_actions, hidden_layers=[300, 1])
#Eself.mu_do_minimize = tf.function(self.mu_minimize, input_signature=[
#tf.TensorSpec(shape=(None, self.num_states), dtype=tf.float32, name='state')])
self.q_do_minimize = tf.function(self.q_minimize, input_signature=[
tf.TensorSpec(shape=(None, self.num_states), dtype=tf.float32, name='state'),
tf.TensorSpec(shape=(None, self.num_actions), dtype=tf.float32, name='action'),
tf.TensorSpec(shape=(None, self.num_states), dtype=tf.float32, name='next_state'),
tf.TensorSpec(shape=(None, ), dtype=tf.float32, name='reward'),
tf.TensorSpec(shape=(None, ), dtype=tf.float32, name='done_flags')])
#tf.function
def train_mu(self, state):
with tf.GradientTape() as tape:
actions = self.mu_model(state, training=True)
critic_value = self.q_model([state, actions], training=True)
# Used `-value` as we want to maximize the value given
# by the critic for our actions
actor_loss = -tf.math.reduce_mean(critic_value)
actor_grad = tape.gradient(actor_loss, self.mu_model.trainable_variables)
self.mu_optimizer.apply_gradients(
zip(actor_grad, self.mu_model.trainable_variables)
)
def q_minimize(self, state, action, next_state, reward, done):
def calc_loss():
q_targ = reward + self.gamma * (1 - done) * self.q_target_model([next_state, action])
q = self.q_model([state, action])
cost = tf.reduce_mean((q - q_targ)**2)
return cost
self.q_optimizer.minimize(calc_loss, self.q_model.trainable_variables)
def train(self, state, action, reward, done, next_state):
state = np.atleast_2d(state)
next_state = np.atleast_2d(next_state)
action = np.atleast_2d(action)
reward = np.atleast_1d(reward)
done = np.atleast_1d(done)
self.update_target_net()
self.train_mu(state)
self.q_do_minimize(state, action, next_state, reward, done)
def update_target_net(self):
mu_weights = np.array(self.mu_model.get_weights())
q_weights = np.array(self.q_model.get_weights())
#print(mu_weights.shape)
#print(q_weights.shape)
mu_target_weights = np.array(self.mu_target_model.get_weights())
q_target_weights = np.array(self.q_target_model.get_weights())
self.q_target_model.set_weights(self.decay * q_weights + (1 - self.decay) * q_target_weights)
def get_action(self, states, noise=None):
if noise is None: noise = self.ACT_NOISE_SCALE
if len(states.shape) == 1: states = states.reshape(1,-1)
action = self.mu_model.predict_on_batch(states)[0]
if noise != 0:
action += noise * np.random.randn(self.num_actions)
action = np.clip(action, -self.action_max, self.action_max)
return action
def play_one(env, agent, replay_buffer, gamma=0.99, noise=0.1, max_episode_len=1000, start_steps=10000, num_train_ep=100, batch_size=100, test_ep_agent=25):
returns = []
num_steps = 0
for ep in range(num_train_ep):
s, ep_return, ep_len, d = env.reset(), 0, 0, False
while not (d or ep_len == max_episode_len):
env.render()
if num_steps > start_steps:
a = agent.get_action(s, noise)
else:
a = env.action_space.sample()
num_steps+=1
if num_steps == start_steps:
print("USING AGENT ACTIONS NOW")
s2, r, d, _ = env.step(a)
ep_return+=r
ep_len+=1
#print(s.shape)
d = False if ep_len == max_episode_len else d
replay_buffer.add_experience(s, a, r, s2, d)
s = s2
for _ in range(ep_len):
batch = replay_buffer.sample_batch()
state, next_state, action, reward, done = batch['s'], batch['s2'], batch['a'], batch['r'], batch['d']
loss = agent.train(state, action, reward, done, next_state)
returns.append(ep_return)
print('Iter:', ep, 'Rewards:', ep_return)
return returns
if __name__ == '__main__':
env = gym.make('Pendulum-v0')
obs_dim1 = env.observation_space.shape[0]
act_dim1 = env.action_space.shape[0]
action_max1 = env.action_space.high[0]
actor = DDPG(env, obs_dim1, act_dim1, action_max1)
replay_buffer = ReplayBuffer(obs_dim1, act_dim1, size=100000)
returns = play_one(env, actor, replay_buffer)
Thanks you in advance!
First things that comes to mind is the learning rate: 0.01 is too high, even for pendulum. Try a lower learning rate (eg 1e-3 for the actor and 5e-3 for the critic).
Also a couple of things look off in your code:
There is no target network for the actor. Why is that? IIRC ddpg has target network for both actor and critic.
Usually it is better to initialize main and target network with the same parameters. You can do that with target_model.set_weights(model.get_weights())
In the function play_one the training steps are done after playing a whole episode. This is probably ok, but there is no need to: because pendulum is not real time you don't need your code to be fast, so you can train while playing.
If you want to take a look I implemented ddpg in tensorflow 2 a while back. It solves pendulum in 80ish episodes.
GitHub
I save the trained model after a certain number of episodes with the special save() function of the DDPG class (the network is saved when the reward reaches zero), but when I restore the model again using saver.restore(), the network gives out a reward equal to approximately -1800. Why is this happening, maybe I'm doing something wrong? My network:
import tensorflow as tf
import numpy as np
import gym
epsiode_steps = 500
# learning rate for actor
lr_a = 0.001
# learning rate for critic
lr_c = 0.002
gamma = 0.9
alpha = 0.01
memory = 10000
batch_size = 32
render = True
class DDPG(object):
def __init__(self, no_of_actions, no_of_states, a_bound, ):
self.memory = np.zeros((memory, no_of_states * 2 + no_of_actions + 1), dtype=np.float32)
# initialize pointer to point to our experience buffer
self.pointer = 0
self.sess = tf.Session()
self.noise_variance = 3.0
self.no_of_actions, self.no_of_states, self.a_bound = no_of_actions, no_of_states, a_bound,
self.state = tf.placeholder(tf.float32, [None, no_of_states], 's')
self.next_state = tf.placeholder(tf.float32, [None, no_of_states], 's_')
self.reward = tf.placeholder(tf.float32, [None, 1], 'r')
with tf.variable_scope('Actor'):
self.a = self.build_actor_network(self.state, scope='eval', trainable=True)
a_ = self.build_actor_network(self.next_state, scope='target', trainable=False)
with tf.variable_scope('Critic'):
q = self.build_crtic_network(self.state, self.a, scope='eval', trainable=True)
q_ = self.build_crtic_network(self.next_state, a_, scope='target', trainable=False)
self.ae_params = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope='Actor/eval')
self.at_params = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope='Actor/target')
self.ce_params = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope='Critic/eval')
self.ct_params = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope='Critic/target')
# update target value
self.soft_replace = [
[tf.assign(at, (1 - alpha) * at + alpha * ae), tf.assign(ct, (1 - alpha) * ct + alpha * ce)]
for at, ae, ct, ce in zip(self.at_params, self.ae_params, self.ct_params, self.ce_params)]
q_target = self.reward + gamma * q_
td_error = tf.losses.mean_squared_error(labels=(self.reward + gamma * q_), predictions=q)
self.ctrain = tf.train.AdamOptimizer(lr_c).minimize(td_error, name="adam-ink", var_list=self.ce_params)
a_loss = - tf.reduce_mean(q)
# train the actor network with adam optimizer for minimizing the loss
self.atrain = tf.train.AdamOptimizer(lr_a).minimize(a_loss, var_list=self.ae_params)
tf.summary.FileWriter("logs2", self.sess.graph)
# initialize all variables
self.sess.run(tf.global_variables_initializer())
self.saver = tf.train.Saver()
self.saver.restore(self.sess, "Pendulum/nn.ckpt")
def choose_action(self, s):
a = self.sess.run(self.a, {self.state: s[np.newaxis, :]})[0]
a = np.clip(np.random.normal(a, self.noise_variance), -2, 2)
return a
def learn(self):
# soft target replacement
self.sess.run(self.soft_replace)
indices = np.random.choice(memory, size=batch_size)
batch_transition = self.memory[indices, :]
batch_states = batch_transition[:, :self.no_of_states]
batch_actions = batch_transition[:, self.no_of_states: self.no_of_states + self.no_of_actions]
batch_rewards = batch_transition[:, -self.no_of_states - 1: -self.no_of_states]
batch_next_state = batch_transition[:, -self.no_of_states:]
self.sess.run(self.atrain, {self.state: batch_states})
self.sess.run(self.ctrain, {self.state: batch_states, self.a: batch_actions, self.reward: batch_rewards,
self.next_state: batch_next_state})
# we define a function store_transition which stores all the transition information in the buffer
def store_transition(self, s, a, r, s_):
trans = np.hstack((s, a, [r], s_))
index = self.pointer % memory
self.memory[index, :] = trans
self.pointer += 1
if self.pointer > memory:
self.noise_variance *= 0.99995
self.learn()
# we define the function build_actor_network for builing our actor network and after crtic network
def build_actor_network(self, s, scope, trainable)
with tf.variable_scope(scope):
l1 = tf.layers.dense(s, 30, activation=tf.nn.tanh, name='l1', trainable=trainable)
a = tf.layers.dense(l1, self.no_of_actions, activation=tf.nn.tanh, name='a', trainable=trainable)
return tf.multiply(a, self.a_bound, name="scaled_a")
def build_crtic_network(self, s, a, scope, trainable):
with tf.variable_scope(scope):
n_l1 = 30
w1_s = tf.get_variable('w1_s', [self.no_of_states, n_l1], trainable=trainable)
w1_a = tf.get_variable('w1_a', [self.no_of_actions, n_l1], trainable=trainable)
b1 = tf.get_variable('b1', [1, n_l1], trainable=trainable)
net = tf.nn.tanh(tf.matmul(s, w1_s) + tf.matmul(a, w1_a) + b1)
q = tf.layers.dense(net, 1, trainable=trainable)
return q
def save(self):
self.saver.save(self.sess, "Pendulum/nn.ckpt")
env = gym.make("Pendulum-v0")
env = env.unwrapped
env.seed(1)
no_of_states = env.observation_space.shape[0]
no_of_actions = env.action_space.shape[0]
a_bound = env.action_space.high
ddpg = DDPG(no_of_actions, no_of_states, a_bound)
total_reward = []
no_of_episodes = 300
# for each episodes
for i in range(no_of_episodes):
# initialize the environment
s = env.reset()
# episodic reward
ep_reward = 0
for j in range(epsiode_steps):
env.render()
# select action by adding noise through OU process
a = ddpg.choose_action(s)
# peform the action and move to the next state s
s_, r, done, info = env.step(a)
# store the the transition to our experience buffer
# sample some minibatch of experience and train the network
ddpg.store_transition(s, a, r, s_)
# update current state as next state
s = s_
# add episodic rewards
ep_reward += r
if int(ep_reward) == 0 and i > 200:
ddpg.save()
print("save")
quit()
if j == epsiode_steps - 1:
total_reward.append(ep_reward)
print('Episode:', i, ' Reward: %i' % int(ep_reward))
break
After finishing Coursera's Practical RL course on A3C, I'm trying to implement my own A3C agent using tensorflow 2. To start, I'm training it on the Cartpole environment but I can't get good results. For now, I've already launched several training with the following code, changing the entropy coefficient to see its impact (the results are shown below). Does it come from my implementation, or is it more a fine-tuning issue ?
class A3C:
def __init__(self, state_dim, n_actions, optimizer=tf.keras.optimizers.Adam(1e-3)):
self.state_input = Input(shape=state_dim)
self.x = Dense(256, activation='relu')(self.state_input)
self.head_v = Dense(1, activation='linear')(self.x)
self.head_p = Dense(n_actions, activation='linear')(self.x)
self.network = tf.keras.Model(inputs=[self.state_input], outputs=[self.head_v, self.head_p])
self.optimizer = optimizer
def forward(self, state):
return self.network(state)
def sample(self, logits):
policy = np.exp(logits.numpy()) / np.sum(np.exp(logits.numpy()), axis=-1, keepdims=True)
return np.array([np.random.choice(len(p), p=p) for p in policy])
def evaluate(agent, env, n_games=1): """Plays an a game from start till done, returns per-game rewards """
game_rewards = []
for _ in range(n_games):
state = env.reset()
total_reward = 0
while True:
action = agent.sample(agent.forward(np.array([state]))[1])[0]
state, reward, done, info = env.step(action)
total_reward += reward
if done: break
game_rewards.append(total_reward)
return game_rewards
class EnvBatch:
def __init__(self, n_envs = 10):
self.envs = [gym.make(env_id) for _ in range(n_envs)]
def reset(self):
return np.array([env.reset() for env in self.envs])
def step(self, actions):
results = [env.step(a) for env, a in zip(self.envs, actions)]
new_obs, rewards, done, infos = map(np.array, zip(*results))
for i in range(len(self.envs)):
if done[i]:
new_obs[i] = self.envs[i].reset()
return new_obs, rewards, done, infos
env_id = "CartPole-v0"
env = gym.make(env_id)
state_dim = env.observation_space.shape
n_actions = env.action_space.n
agent = A3C(state_dim, n_actions)
env_batch = EnvBatch(10)
batch_states = env_batch.reset()
gamma=0.99
rewards_history = []
entropy_history = []
for i in trange(200000):
with tf.GradientTape() as t:
batch_values, batch_logits = agent.forward(batch_states)
batch_actions = agent.sample(batch_logits)
batch_next_states, batch_rewards, batch_dones, _ = env_batch.step(batch_actions)
batch_next_values, btach_next_logits = agent.forward(batch_next_states)
batch_next_values *= (1 - batch_dones)
probs = tf.nn.softmax(batch_logits)
logprobs = tf.nn.log_softmax(batch_logits)
logp_actions = tf.reduce_sum(logprobs * tf.one_hot(batch_actions, n_actions), axis=-1)
advantage = batch_rewards + gamma*batch_next_values - batch_values
entropy = -tf.reduce_sum(probs * logprobs, 1, name="entropy")
actor_loss = - tf.reduce_mean(logp_actions * tf.stop_gradient(advantage)) - 0.005 * tf.reduce_mean(entropy)
target_state_values = batch_rewards + gamma*batch_next_values
critic_loss = tf.reduce_mean((batch_values - tf.stop_gradient(target_state_values))**2 )
loss = actor_loss + critic_loss
var_list = agent.network.trainable_variables
grads = t.gradient(loss,var_list)
agent.optimizer.apply_gradients(zip(grads, var_list))
batch_states = batch_next_states
entropy_history.append(np.mean(entropy))
if i % 500 == 0:
if i % 2500 == 0:
rewards_history.append(np.mean(evaluate(agent, env, n_games=3)))
clear_output(True)
plt.figure(figsize=[8, 4])
plt.subplot(1, 2, 1)
plt.plot(rewards_history, label='rewards')
plt.title("Session rewards")
plt.grid()
plt.legend()
plt.subplot(1, 2, 2)
plt.plot(entropy_history, label='entropy')
plt.title("Policy entropy")
plt.grid()
plt.legend()
plt.show()
Beta = 0.005 - Training 1
Beta = 0.005 - Training 2
Beta = 0.005 - Training 3
Beta = 0.05 - Training 1
Beta = 0.05 - Training 2
Beta = 0.05 - Training 3
I've looked through your code, and it doesn't look like there's any problem with the algorithm. That is, it seems to me that the Hyper Parameter was chosen incorrectly. Try different Hyper Parameter Sets. If it doesn't work properly, refer to repository
The critic loss is wrong. You should get first expect returns, predicting the next state and iterate over it with bellman equation.
Here is an example:
def getExpectedReturns(self, states, next_states, done, rewards, standarize=True):
# Get next value
if done[-1] == 1.0:
arr_idx = np.zeros((rewards.shape[0], 1))
arr_idx[-1] = 1.0
values_rewards_sum_one_hot = tf.convert_to_tensor(arr_idx, dtype=tf.float32)
next_value = tf.reduce_sum(rewards * values_rewards_sum_one_hot, axis=0)
else:
values_rewards_sum = self.model_a2c(next_states)[-1]
arr_idx = np.zeros((rewards.shape[0], 1))
arr_idx[0] = 1.0
values_rewards_sum_one_hot = tf.convert_to_tensor(arr_idx, dtype=tf.float32)
next_value = tf.reduce_sum(values_rewards_sum * values_rewards_sum_one_hot, axis=0)
# Iterate over rewards
list_true_values = []
for i in reversed(range(0, len(rewards))):
if done[i]==0.0:
next_value = rewards[i] + next_value * self.gamma
else:
next_value = rewards[i]
list_true_values.append(next_value)
list_true_values.reverse()
list_true_values = tf.convert_to_tensor(list_true_values, dtype=tf.float32)
if standarize:
list_true_values = ((list_true_values - tf.math.reduce_mean(list_true_values)) /
(tf.math.reduce_std(list_true_values) + tf.constant(1e-12)))
return list_true_values
with tf.GradientTape() as tape:
# Advantage
returns = self.getExpectedReturns(states, next_states, done, rewards, standarize=False)
actions_probs_logits, values = self.model_a2c(states)
advantage = returns - values
advantage = tf.squeeze(advantage)
# Actions probs
actions_probs_softmax = tf.nn.softmax(actions_probs_logits)
actions_log_probs_softmax = tf.nn.log_softmax(actions_probs_logits)
actions_one_hot = tf.one_hot(actions, self.num_actions, 1.0, 0.0)
actions_log_probs = tf.reduce_sum(actions_log_probs_softmax * actions_one_hot, axis=-1)
# Entropy
entropy = self.entropy_coef * tf.reduce_mean(actions_probs_softmax * actions_log_probs_softmax, axis=1)
# Losses
actor_loss = -tf.reduce_mean(actions_log_probs * tf.stop_gradient(advantage), axis=0)
critic_loss = self.critic_coef * tf.reduce_mean(tf.math.pow(advantage, 2), axis=0)
total_loss = actor_loss + critic_loss - entropy
I have two python files namely lstm.py and test.py in same folder. im executing the test.py file in python interactive shell. It shows the
AttributeError: Lstm Param instance has no attribute "x_list_add".
lstm.py
import random
import numpy as np
import math
def sigmoid(x):
return 1. / (1 + np.exp(-x))
def sigmoid_derivative(values):
return values*(1-values)
def tanh_derivative(values):
return 1. - values ** 2
# createst uniform random array w/ values in [a,b) and shape args
def rand_arr(a, b, *args):
np.random.seed(0)
return np.random.rand(*args) * (b - a) + a
class LstmParam:
def __init__(self, mem_cell_ct, x_dim):
self.mem_cell_ct = mem_cell_ct
self.x_dim = x_dim
concat_len = x_dim + mem_cell_ct
# weight matrices
self.wg = rand_arr(-0.1, 0.1, mem_cell_ct, concat_len)
self.wi = rand_arr(-0.1, 0.1, mem_cell_ct, concat_len)
self.wf = rand_arr(-0.1, 0.1, mem_cell_ct, concat_len)
self.wo = rand_arr(-0.1, 0.1, mem_cell_ct, concat_len)
# bias terms
self.bg = rand_arr(-0.1, 0.1, mem_cell_ct)
self.bi = rand_arr(-0.1, 0.1, mem_cell_ct)
self.bf = rand_arr(-0.1, 0.1, mem_cell_ct)
self.bo = rand_arr(-0.1, 0.1, mem_cell_ct)
# diffs (derivative of loss function w.r.t. all parameters)
self.wg_diff = np.zeros((mem_cell_ct, concat_len))
self.wi_diff = np.zeros((mem_cell_ct, concat_len))
self.wf_diff = np.zeros((mem_cell_ct, concat_len))
self.wo_diff = np.zeros((mem_cell_ct, concat_len))
self.bg_diff = np.zeros(mem_cell_ct)
self.bi_diff = np.zeros(mem_cell_ct)
self.bf_diff = np.zeros(mem_cell_ct)
self.bo_diff = np.zeros(mem_cell_ct)
def apply_diff(self, lr = 1):
self.wg -= lr * self.wg_diff
self.wi -= lr * self.wi_diff
self.wf -= lr * self.wf_diff
self.wo -= lr * self.wo_diff
self.bg -= lr * self.bg_diff
self.bi -= lr * self.bi_diff
self.bf -= lr * self.bf_diff
self.bo -= lr * self.bo_diff
# reset diffs to zero
self.wg_diff = np.zeros_like(self.wg)
self.wi_diff = np.zeros_like(self.wi)
self.wf_diff = np.zeros_like(self.wf)
self.wo_diff = np.zeros_like(self.wo)
self.bg_diff = np.zeros_like(self.bg)
self.bi_diff = np.zeros_like(self.bi)
self.bf_diff = np.zeros_like(self.bf)
self.bo_diff = np.zeros_like(self.bo)
class LstmState:
def __init__(self, mem_cell_ct, x_dim):
self.g = np.zeros(mem_cell_ct)
self.i = np.zeros(mem_cell_ct)
self.f = np.zeros(mem_cell_ct)
self.o = np.zeros(mem_cell_ct)
self.s = np.zeros(mem_cell_ct)
self.h = np.zeros(mem_cell_ct)
self.bottom_diff_h = np.zeros_like(self.h)
self.bottom_diff_s = np.zeros_like(self.s)
class LstmNode:
def __init__(self, lstm_param, lstm_state):
# store reference to parameters and to activations
self.state = lstm_state
self.param = lstm_param
# non-recurrent input concatenated with recurrent input
self.xc = None
def bottom_data_is(self, x, s_prev = None, h_prev = None):
# if this is the first lstm node in the network
if s_prev is None: s_prev = np.zeros_like(self.state.s)
if h_prev is None: h_prev = np.zeros_like(self.state.h)
# save data for use in backprop
self.s_prev = s_prev
self.h_prev = h_prev
# concatenate x(t) and h(t-1)
xc = np.hstack((x, h_prev))
self.state.g = np.tanh(np.dot(self.param.wg, xc) + self.param.bg)
self.state.i = sigmoid(np.dot(self.param.wi, xc) + self.param.bi)
self.state.f = sigmoid(np.dot(self.param.wf, xc) + self.param.bf)
self.state.o = sigmoid(np.dot(self.param.wo, xc) + self.param.bo)
self.state.s = self.state.g * self.state.i + s_prev * self.state.f
self.state.h = self.state.s * self.state.o
self.xc = xc
def top_diff_is(self, top_diff_h, top_diff_s):
# notice that top_diff_s is carried along the constant error carousel
ds = self.state.o * top_diff_h + top_diff_s
do = self.state.s * top_diff_h
di = self.state.g * ds
dg = self.state.i * ds
df = self.s_prev * ds
# diffs w.r.t. vector inside sigma / tanh function
di_input = sigmoid_derivative(self.state.i) * di
df_input = sigmoid_derivative(self.state.f) * df
do_input = sigmoid_derivative(self.state.o) * do
dg_input = tanh_derivative(self.state.g) * dg
# diffs w.r.t. inputs
self.param.wi_diff += np.outer(di_input, self.xc)
self.param.wf_diff += np.outer(df_input, self.xc)
self.param.wo_diff += np.outer(do_input, self.xc)
self.param.wg_diff += np.outer(dg_input, self.xc)
self.param.bi_diff += di_input
self.param.bf_diff += df_input
self.param.bo_diff += do_input
self.param.bg_diff += dg_input
# compute bottom diff
dxc = np.zeros_like(self.xc)
dxc += np.dot(self.param.wi.T, di_input)
dxc += np.dot(self.param.wf.T, df_input)
dxc += np.dot(self.param.wo.T, do_input)
dxc += np.dot(self.param.wg.T, dg_input)
# save bottom diffs
self.state.bottom_diff_s = ds * self.state.f
self.state.bottom_diff_h = dxc[self.param.x_dim:]
class LstmNetwork():
def __init__(self, lstm_param):
self.lstm_param = lstm_param
self.lstm_node_list = []
# input sequence
self.x_list = []
def y_list_is(self, y_list, loss_layer):
"""
Updates diffs by setting target sequence
with corresponding loss layer.
Will *NOT* update parameters. To update parameters,
call self.lstm_param.apply_diff()
"""
assert len(y_list) == len(self.x_list)
idx = len(self.x_list) - 1
# first node only gets diffs from label ...
loss = loss_layer.loss(self.lstm_node_list[idx].state.h, y_list[idx])
diff_h = loss_layer.bottom_diff(self.lstm_node_list[idx].state.h, y_list[idx])
# here s is not affecting loss due to h(t+1), hence we set equal to zero
diff_s = np.zeros(self.lstm_param.mem_cell_ct)
self.lstm_node_list[idx].top_diff_is(diff_h, diff_s)
idx -= 1
### ... following nodes also get diffs from next nodes, hence we add diffs to diff_h
### we also propagate error along constant error carousel using diff_s
while idx >= 0:
loss += loss_layer.loss(self.lstm_node_list[idx].state.h, y_list[idx])
diff_h = loss_layer.bottom_diff(self.lstm_node_list[idx].state.h, y_list[idx])
diff_h += self.lstm_node_list[idx + 1].state.bottom_diff_h
diff_s = self.lstm_node_list[idx + 1].state.bottom_diff_s
self.lstm_node_list[idx].top_diff_is(diff_h, diff_s)
idx -= 1
return loss
def x_list_clear(self):
self.x_list = []
def x_list_add(self, x):
self.x_list.append(x)
if len(self.x_list) > len(self.lstm_node_list):
# need to add new lstm node, create new state mem
lstm_state = LstmState(self.lstm_param.mem_cell_ct, self.lstm_param.x_dim)
self.lstm_node_list.append(LstmNode(self.lstm_param, lstm_state))
# get index of most recent x input
idx = len(self.x_list) - 1
if idx == 0:
# no recurrent inputs yet
self.lstm_node_list[idx].bottom_data_is(x)
else:
s_prev = self.lstm_node_list[idx - 1].state.s
h_prev = self.lstm_node_list[idx - 1].state.h
self.lstm_node_list[idx].bottom_data_is(x, s_prev, h_prev)
test.py
import numpy as np
import sys
sys.path.append("/home/vj/Downloads/tfdeep/core_lstm/lstm/")
from lstm import LstmParam, LstmNetwork
class ToyLossLayer:
"""
Computes square loss with first element of hidden layer array."""
#classmethod
def loss(self, pred, label):
return (pred[0] - label) ** 2
#classmethod
def bottom_diff(self, pred, label):
diff = np.zeros_like(pred)
diff[0] = 2 * (pred[0] - label)
return diff
def example_0():
# learns to repeat simple sequence from random inputs
np.random.seed(0)
# parameters for input data dimension and lstm cell count
mem_cell_ct = 100
x_dim = 50
lstm_param = LstmParam(mem_cell_ct, x_dim)
lstm_net = LstmNetwork(lstm_param)
y_list = [-0.5, 0.2, 0.1, -0.5]
input_val_arr = [np.random.random(x_dim) for _ in y_list]
for cur_iter in range(100):
print("iter", "%2s" % str(cur_iter),": ")
for ind in range(len(y_list)):
lstm_net.x_list_add(input_val_arr[ind])
#print("y_pred = [" + ", ".join(["% 2.5f" % lstm_net.lstm_node_list[ind].state.h[0] for ind in range(len(y_list))]) +"]," end=":")
loss = lstm_net.y_list_is(y_list, ToyLossLayer)
print("loss:", "%.3e" % loss)
lstm_param.apply_diff(lr=0.1)
lstm_net.x_list_clear()
if __name__ == "__main__":
example_0()
while executing this test.py file I am getting I getting the follwing error
Traceback (most recent call last):
File "", line 2, in
File "", line 14, in example_0
AttributeError: LstmNetwork instance has no attribute 'x_list_add'
I am struggling to get the output. I didnot get any post relate to my problem.
tl;dr: I input a word to my model, and am supposed to get a list of similar words and their associated measures of similarity back. I get an error: Aborted (core dumped).
My goal is to determine which words are similar to an input word, based on their feature vectors. I have model already trained. I load it and call two functions:
def main(argv=None):
model = NVDM(args)
sess_saver = tf.train.Saver()
sess = tf.Session()
init = tf.initialize_all_variables()
sess.run(init)
loaded = load_for_similar(sess, sess_saver) #my function
wm = word_match(sess, loaded[0], loaded[1], "bottle", loaded[2], loaded[3], topN=5)
My problem is that I can't print out the words which are similar and the associated similarity measure. I tried (in main):
sess.run(wm)
wm[0].eval(session=sess)
print(wm)
All of which gave me the error:
F tensorflow/core/kernels/strided_slice_op.cc:316] Check failed: tmp.CopyFrom(input.Slice(begin[0], end[0]), final_shape)
Aborted (core dumped)
This tells me I'm not running the session properly. What am I doing wrong?
Details on the functions, just in case:
The function 'load_for_similar' restores the weights and bias of the decoder in my model (a variational autoencoder), and normalizes them. It also reverses the order of the keys and values in my vocabulary dictionary for later use:
def load_for_similar(sess, saver_obj):
saver_obj.restore(sess, "./CA_checkpoints/saved_model.ckpt")
vocab_file = '/path/to/vocab.pkl'
t1 = loader_object(vocab_file)
v1 = t1.get_vocab()
v1_rev = {k:v for v, k in v1.iteritems()}
decoder_mat = tf.get_collection(tf.GraphKeys.VARIABLES, scope='decoder')[0]
decoder_bias = tf.get_collection(tf.GraphKeys.VARIABLES, scope='decoder')[1]
return (find_norm(decoder_mat), find_norm(decoder_bias), v1, v1_rev)
To find similar words, I pass the normalized weight matrix and bias in to an new function, along with the feature vector of my word (vec):
def find_similar(sess, Weights, vec, bias):
dists = tf.add(tf.reduce_sum(tf.mul(Weights, vec)), bias)
best = argsort(sess, dists, reverse=True)
dist_sort = tf.nn.top_k(dists, k=dists.get_shape().as_list()[0], sorted=True).values
return dist_sort, best
Finally, I want to match the words that are closest to my supplied word, "bottle":
def word_match(sess, norm_mat , norm_bias, word_ , vocab, vocab_inverse , topN = 10):
idx = vocab[word_]
similarity_meas , indexes = find_similar(sess, norm_mat , norm_mat[idx], norm_bias)
words = tf.gather(vocab_inverse.keys(), indexes[:topN])
return (words, similarity_meas[:topN])
EDIT: in response to mrry's comment, here is the model (I hope this is what you wanted?). This code depends on utils.py, a separate utilities file. I will include that as well. Please note that this code is heavily based on Yishu Miao's and Sarath Nair's.
class NVDM(object):
""" Neural Variational Document Model -- BOW VAE.
"""
def __init__(self,
vocab_size=15000, #was 2000
n_hidden=500,
n_topic=50,
n_sample=1,
learning_rate=1e-5,
batch_size=100, #was 64
non_linearity=tf.nn.tanh):
self.vocab_size = vocab_size
self.n_hidden = n_hidden
self.n_topic = n_topic
self.n_sample = n_sample
self.non_linearity = non_linearity
self.learning_rate = learning_rate/batch_size #CA
self.batch_size = batch_size
self.x = tf.placeholder(tf.float32, [None, vocab_size], name='input')
self.mask = tf.placeholder(tf.float32, [None], name='mask') # mask paddings
# encoder
with tf.variable_scope('encoder'):
self.enc_vec = utils.mlp(self.x, [self.n_hidden, self.n_hidden])
self.mean = utils.linear(self.enc_vec, self.n_topic, scope='mean')
self.logsigm = utils.linear(self.enc_vec,
self.n_topic,
bias_start_zero=True,
matrix_start_zero=False,
scope='logsigm')
self.kld = -0.5 * tf.reduce_sum(1 - tf.square(self.mean) + 2 * self.logsigm - tf.exp(2 * self.logsigm), 1)
self.kld = self.mask*self.kld # mask paddings
with tf.variable_scope('decoder'):
if self.n_sample ==1: # single sample
p1 = tf.cast(tf.reduce_sum(self.mask), tf.int32) #needed for random normal generation
eps = tf.random_normal((p1, self.n_topic), 0, 1)
doc_vec = tf.mul(tf.exp(self.logsigm), eps) + self.mean
logits = tf.nn.log_softmax(utils.linear(doc_vec, self.vocab_size, scope='projection'))
self.recons_loss = -tf.reduce_sum(tf.mul(logits, self.x), 1)
# multiple samples
else:
eps = tf.random_normal((self.n_sample*batch_size, self.n_topic), 0, 1)
eps_list = tf.split(0, self.n_sample, eps)
recons_loss_list = []
for i in xrange(self.n_sample):
if i > 0: tf.get_variable_scope().reuse_variables()
curr_eps = eps_list[i]
doc_vec = tf.mul(tf.exp(self.logsigm), curr_eps) + self.mean
logits = tf.nn.log_softmax(utils.linear(doc_vec, self.vocab_size, scope='projection'))
recons_loss_list.append(-tf.reduce_sum(tf.mul(logits, self.x), 1))
self.recons_loss = tf.add_n(recons_loss_list) / self.n_sample
self.objective = self.recons_loss + self.kld
optimizer = tf.train.AdamOptimizer(learning_rate=self.learning_rate)
fullvars = tf.trainable_variables()
enc_vars = utils.variable_parser(fullvars, 'encoder')
dec_vars = utils.variable_parser(fullvars, 'decoder')
enc_grads = tf.gradients(self.objective, enc_vars)
dec_grads = tf.gradients(self.objective, dec_vars)
self.optim_enc = optimizer.apply_gradients(zip(enc_grads, enc_vars))
self.optim_dec = optimizer.apply_gradients(zip(dec_grads, dec_vars))
def minibatch_bow(it1, Instance1, n_samples, batch_size, used_ints = set()):
available = set(np.arange(n_samples)) - used_ints #
if len(available) < batch_size:
indices = np.array(list(available))
else:
indices = np.random.choice(tuple(available), batch_size, replace=False)
used = used_ints
mb = itemgetter(*indices)(it1)
batch_xs = Instance1._bag_of_words(mb, vocab_size=15000)
batch_flattened = np.ravel(batch_xs)
index_positions = np.where(batch_flattened > 0)[0]
return (batch_xs, index_positions, set(indices)) #batch_xs[0] is the bag of words; batch_xs[1] is the 0/1 word used/not;
def train(sess, model, train_file, vocab_file, saver_obj, training_epochs, alternate_epochs, batch_size):
Instance1 = testchunk_Nov23.testLoader(train_file, vocab_file)
data_set = Instance1.get_batch(batch_size) #get all minibatches of size 100
n_samples = Instance1.num_reviews()
train_batches = list(data_set) #this is an itertools.chain object
it1_train = list(itertools.chain(*train_batches)) #length is 732,356. This is all the reviews.atch_size
if len(it1_train) % batch_size != 0:
total_batch = int(len(it1_train)/batch_size) + 1
else:
total_batch = int(len(it1_train)/batch_size)
trainfilesave = "train_ELBO_and_perplexity_Dec1.txt"
#Training
train_time = time.time()
for epoch in range(training_epochs):
for switch in xrange(0, 2):
if switch == 0:
optim = model.optim_dec
print_mode = 'updating decoder'
else:
optim = model.optim_enc
print_mode = 'updating encoder'
with open(trainfilesave, 'w') as f:
for i in xrange(alternate_epochs):
loss_sum = 0.0
kld_sum = 0.0
word_count = 0
used_indices = set()
for idx_batch in range(total_batch): #train_batches:
mb = minibatch_bow(it1_train, Instance1, n_samples, batch_size, used_ints=used_indices)
print('minibatch', idx_batch)
used_indices.update(mb[2])
num_mb = np.ones(mb[0][0].shape[0])
input_feed = {model.x.name: mb[0][0], model.mask: num_mb}
_, (loss, kld) = sess.run((optim,[model.objective, model.kld]) , input_feed)
loss_sum += np.sum(loss)
And the utils.py file:
def linear(inputs,
output_size,
no_bias=False,
bias_start_zero=False,
matrix_start_zero=False,
scope=None):
"""Define a linear connection."""
with tf.variable_scope(scope or 'Linear'):
if matrix_start_zero:
matrix_initializer = tf.constant_initializer(0)
else:
matrix_initializer = None
if bias_start_zero:
bias_initializer = tf.constant_initializer(0)
else:
bias_initializer = None
input_size = inputs.get_shape()[1].value
matrix = tf.get_variable('Matrix', [input_size, output_size],
initializer=matrix_initializer)
bias_term = tf.get_variable('Bias', [output_size],
initializer=bias_initializer)
output = tf.matmul(inputs, matrix)
if not no_bias:
output = output + bias_term
return output
def mlp(inputs,
mlp_hidden=[],
mlp_nonlinearity=tf.nn.tanh,
scope=None):
"""Define an MLP."""
with tf.variable_scope(scope or 'Linear'):
mlp_layer = len(mlp_hidden)
res = inputs
for l in xrange(mlp_layer):
res = mlp_nonlinearity(linear(res, mlp_hidden[l], scope='l'+str(l)))
return res