-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathjustfile
70 lines (57 loc) · 2.36 KB
/
justfile
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
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
OPTIONS := "lim_malek"
agent := "dsup"
seed := "0"
num_steps := "200000"
option_idx := "0"
env := OPTIONS
DQN := "config.agents.dqn"
DAU := "config.agents.dau"
QRDQN := "config.agents.qrdqn"
DSUP := "config.agents.dsup"
agent_config := "config.agents." + agent
EVAL_EPISODES_OPTIONS := "200"
writer := "aim"
time_mul := "1"
risk_type := "neutral"
risk_param := "1.0"
WORKDIR := "models" / env / agent / "h-" + time_mul + "_risk-" + risk_type + "-" + risk_param + "_v" + seed
train agent env eval_episodes:
python -m dsup.testbed \
--agent config:{{agent_config}}.{{env}} \
--agent set:seed={{seed}} \
--agent set:num_steps={{num_steps}} \
--agent set:warmup_steps=0 \
--agent fiddler:config.fiddlers.set_decision_frequency\(decision_frequency={{time_mul}}\) \
--experiment {{env}} \
--metric_writer {{writer}} \
--workdir {{WORKDIR}} \
--num_eval_episodes {{eval_episodes}}
train_dsup_shifted agent env eval_episodes:
python -m dsup.testbed \
--agent config:{{agent_config}}.{{env}} \
--agent set:seed={{seed}} \
--agent set:num_steps={{num_steps}} \
--agent set:warmup_steps=0 \
--agent fiddler:config.fiddlers.set_decision_frequency\(decision_frequency={{time_mul}}\) \
--agent fiddler:config.fiddlers.dsup_shifted\(rescale_factor=0.5\) \
--experiment {{env}} \
--metric_writer {{writer}} \
--workdir {{WORKDIR}} \
--enable_checkpoints \
--num_eval_episodes {{eval_episodes}}
OPTION_IDX := OPTIONS + "\\(idx=" + option_idx + "\\)"
train_options: (train agent OPTION_IDX EVAL_EPISODES_OPTIONS)
train_options_dsup_shifted: (train_dsup_shifted agent OPTION_IDX EVAL_EPISODES_OPTIONS)
train_risky agent env eval_episodes:
python -m dsup.testbed \
--agent config:{{agent_config}}.{{env}} \
--agent set:seed={{seed}} \
--agent set:num_steps={{num_steps}} \
--agent fiddler:config.fiddlers.set_decision_frequency\(decision_frequency={{time_mul}}\) \
--agent fiddler:config.fiddlers.set_risk_measure_cvar\({{risk_param}}\) \
--experiment {{env}}-risky \
--metric_writer {{writer}} \
--workdir {{WORKDIR}} \
--enable_checkpoints \
--num_eval_episodes {{eval_episodes}}
train_options_risky: (train_risky agent OPTION_IDX EVAL_EPISODES_OPTIONS)