Hw1/learn - #1

Open

CeHao1 wants to merge 36 commits into

berkeleydeeprlcourse:mainfrom

CeHao1:hw1/learn

Commits on Sep 1, 2021

1
ce-hao
committed
2ed7552
finish two utils
ce-hao
committed
6b0b810
finish the replay sample
ce-hao
committed
f1aaee7
finish the replay sample
ce-hao
committed
deeeef1
checked two utils, they are correct;
ce-hao
committed
2834a91
other students use strange ways in replay buffer
ce-hao
committed
afaa788
ok trainer
ce-hao
committed
832f143
mlp policy get action has many implementations
ce-hao
committed
8f21d47
this seems ok
ce-hao
committed
9b4a41d

Commits on Sep 3, 2021

add pickle
ce-hao
committed
94f5b87
add self.sample_recent_data
ce-hao
committed
a643df7
change sample random
ce-hao
committed
c14d118
change sample random
ce-hao
committed
12b0302
add reset
ce-hao
committed
5b4ad55
this tried to imitate human
ce-hao
committed
523a3d8
action
ce-hao
committed
0ebfc4c
action
ce-hao
committed
3414ea3
action
ce-hao
committed
1267a99
1
ce-hao
committed
46ee849
1
ce-hao
committed
e5b9069
1
ce-hao
committed
4d3fb5f
start the plot frame
ce-hao
committed
46825b2
the plot is ok
ce-hao
committed
9960dc9
the plot is ok
ce-hao
committed
222ffe9
1
ce-hao
committed
6701e7f
1
ce-hao
committed
1af92e4
it can save
ce-hao
committed
6241645
1
ce-hao
committed
f83e622
copy from colab
ce-hao
committed
d658381
start adding more ep_len
ce-hao
committed
be8b18e
to test the expert policy
ce-hao
committed
9736f80
last forgot eval reward
ce-hao
committed
0f2bf5e
this one is good for ploting
ce-hao
committed
e3ebe55
super cool;
ce-hao
committed
535ea36
yes
ce-hao
committed
1130307
removed unuseful files
ce-hao
committed
0e6ba84