Highest Voted 'dqn' Questions

1

vote

1 answer

Convergence time of Q-learning Vs Deep Q-learning

I want to know about the convergence time of Deep Q-learning vs Q-learning when run on same problem. Can anyone give me an idea about the pattern between them? It will be better if it is explained with a graph.

deep-learning q-learning dqn

asked Apr 26 '21 at 06:07

A_Tahsin

31
3

1

vote

0 answers

Model output "Tensor("activation_9/activation_9/Identity:0", shape=(?, 6), dtype=float32)" has invalid shape

I am getting this error when I am trying to build an DQN model but I am getting this error: ValueError Traceback (most recent call last) in () 1 # TODO - Select the…

python tensorflow keras dqn keras-rl

asked Apr 19 '21 at 19:34

Michal Ružička Ružička

1,045
8
25

1

vote

0 answers

Understanding the reward functionality in Reinforcment learning (atari breakout)

I'm trying to understand the reward functionality in Breakout atari implemented by Deepmind. I'm a little confused about the reward. They represent every state using four frames and depending on that the reward for every action will be received…

reinforcement-learning dqn reward

asked Mar 04 '21 at 14:34

jon

11
4

1

vote

0 answers

Keras Double DQN average reward decreases over time and is unable to converge

I am attempting to teach a Double DQN agent to run a gridworld where there is one seeker (the agent) who will try to collect all the hiders which are randomly spawned. Every step has a path_cost of -0.1 and if a hider is collected a reward of 1 is…

python tensorflow keras reinforcement-learning dqn

asked Feb 19 '21 at 07:29

KKastberg

11
2

1

vote

1 answer

How does work this implementation of DQN algorithm on TensorFlowJs?

devs, I found a bunch of examples of DQN implementations, but because I'm no TensorFlow expert, I'm a little bit confused. Let's see here is one one of them. I can understand, on the 73rd line, we slice some batch of stored data [{state, action,…

javascript tensorflow tensorflow.js q-learning dqn

asked Jan 08 '21 at 19:33

Vadim

13
5

1

vote

1 answer

DQN doesn't learn

I'm trying to implement a DQN in CarPole environment using Pytorch. I don't know why, but no matter how long I've tried to train the agent, even though the scores generally increase, they just fluctuate without maintaining high scores. The code was…

python pytorch dqn

asked Dec 21 '20 at 17:53

Minh Luke

11
1

1

vote

0 answers

TF_Agent with observation_spec and action_spec being 2D matrices

I am trying to use tf_agent library with a custom environment to build an agent that does the following: The agent serves a number of users (10 users) by giving them resources (12 resources). It receives an input (observation / state) that is a 2D…

python tensorflow keras deep-learning dqn

asked Nov 18 '20 at 10:08

Ness

158
1
12

1

vote

1 answer

Use TensorBoard to visualize graph from tf_agents

I'm quite new to RL and currently teaching myself how to implement different algorithms and hyper-parameters using tf_agents library. I've been playing around with the code provided from this tutorial…

tensorflow tensorboard reinforcement-learning dqn

asked Nov 12 '20 at 07:01

Kai Yun

97
8

1

vote

1 answer

Input states for Deep Q Learning

I am using the DQN for a resource allocation where the agent should assign the arrival requests to the best Virtual Machine. I am modifying a Cartpole code as follow: import random import gym import numpy as np from collections import deque from…

state reinforcement-learning dqn markov-decision-process

asked Oct 05 '20 at 19:09

Alireza Hosseini

19
3

1

vote

1 answer

DQN doesn't make any progress after a little while

Here is my code, its a simple DQN that learns to play snake and i dont know why it stops learning after a little while, for example. it learns that the snake head should hit the wall, but it doesnt learn to eat the fruit, even though i give a reward…

python tensorflow artificial-intelligence reinforcement-learning dqn

asked Sep 13 '20 at 20:39

user13923578

11
1

1

vote

0 answers

Understanding the behavior of tf.random.seed

I usually set a seed for experiment reproducibility when working with tensorflow. In this case my code is as follows: class DeepNetworkModel(tf.keras.Model): def __init__(self, seed: int, input_shape:…

python tensorflow reinforcement-learning random-seed dqn

asked Jul 30 '20 at 14:22

AleB

153
1
3
10

1

vote

1 answer

DQN: Access to raw observations after conversion of observation space into a Box environment?

I'm adapting the Pytorch code from Tabor's course on DQNs (https://github.com/philtabor/Deep-Q-Learning-Paper-To-Code) to work with the vizdoomgym library, having previously managed to make a version work in TF. After training my agent, I will…

reinforcement-learning openai-gym dqn openai-api

asked Jul 25 '20 at 12:42

Y. Xu

11
3

1

vote

0 answers

Rewards not converging in DQN Model Problem

I was recently following Sentdex’s tutorial on DQN’s (https://pythonprogramming.net/deep-q-learning-dqn-reinforcement-learning-python-tutorial/), where he taught a player to move through an environment and get “food” and I decided that I wanted to…

python keras deep-learning reinforcement-learning dqn

asked Jul 21 '20 at 00:52

K. Kwan

21
1
4

1

vote

1 answer

DDQN cannot solve Tic Tac Toe

I am trying to solve Tic Tac Toe with a DDQN approach since a long time now. It took me a while to fill my holes in knowledge, but now my code seems fine. However, I am unsure about how to train the agent, since it is a two player game. Currently, I…

python deep-learning neural-network reinforcement-learning dqn

asked Jun 25 '20 at 08:01

spadel

998
2
16
40

1

vote

1 answer

DQN performs bad in validation mode

I made a DQN to learn tic tac toe. So far, I let the agent play all moves in order to see, if it learns to always make legal moves that leads to a draw or to a win for one or the other player. After training the network for about 10.000 games, it is…

python deep-learning neural-network reinforcement-learning dqn

asked Jun 22 '20 at 20:07

spadel

998
2
16
40

Questions tagged [dqn]