forked from NeuroCSUT/DeepMind-Atari-Deep-Q-Learner-2Player
-
Notifications
You must be signed in to change notification settings - Fork 0
/
run_cpu
executable file
·47 lines (43 loc) · 1.59 KB
/
run_cpu
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
#!/bin/bash
if [ -z "$1" ]
then echo "Please provide the name of the game, e.g. ./run_cpu breakout "; exit 0
fi
ENV=$1
FRAMEWORK="alewrap"
game_path=$PWD"/roms/"
env_params="useRGB=true"
agent="NeuralQLearner"
n_replay=1
netfile="\"convnet_atari3\""
update_freq=4
actrep=4
discount=0.99
seed=1
learn_start=50000
pool_frms_type="\"max\""
pool_frms_size=2
initial_priority="false"
replay_memory=1000000
eps_end=0.1
eps_endt=replay_memory
lr=0.00025
agent_type="DQN3_0_1"
preproc_net="\"net_downsample_2x_full_y\""
agent_name=$agent_type"_"$1"_FULL_Y"
state_dim=7056
ncols=1
agent_params="lr="$lr",ep=1,ep_end="$eps_end",ep_endt="$eps_endt",discount="$discount",hist_len=4,learn_start="$learn_start",replay_memory="$replay_memory",update_freq="$update_freq",n_replay="$n_replay",network="$netfile",preproc="$preproc_net",state_dim="$state_dim",minibatch_size=32,rescale_r=1,ncols="$ncols",bufferSize=512,valid_size=500,target_q=10000,clip_delta=1,min_reward=-1,max_reward=1"
steps=50000000
eval_freq=250000
eval_steps=125000
prog_freq=5000
save_freq=125000
save_versions=$save_freq
gpu=-1
random_starts=30
pool_frms="type="$pool_frms_type",size="$pool_frms_size
num_threads=4
args="-framework $FRAMEWORK -game_path $game_path -name $agent_name -env $ENV -env_params $env_params -agent $agent -agent_params $agent_params -steps $steps -eval_freq $eval_freq -eval_steps $eval_steps -prog_freq $prog_freq -save_freq $save_freq -save_versions $save_versions -actrep $actrep -gpu $gpu -random_starts $random_starts -pool_frms $pool_frms -seed $seed -threads $num_threads"
echo $args
cd dqn
../torch/bin/qlua train_agent.lua $args