Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
OpenDILab开源决策智能平台
DI-engine
提交
414b5305
D
DI-engine
项目概览
OpenDILab开源决策智能平台
/
DI-engine
上一次同步 接近 3 年
通知
65
Star
322
Fork
1
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
0
列表
看板
标记
里程碑
合并请求
0
DevOps
流水线
流水线任务
计划
Wiki
1
Wiki
分析
仓库
DevOps
项目成员
Pages
D
DI-engine
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
0
Issue
0
列表
看板
标记
里程碑
合并请求
0
合并请求
0
Pages
DevOps
DevOps
流水线
流水线任务
计划
分析
分析
仓库分析
DevOps
Wiki
1
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
流水线任务
提交
Issue看板
体验新版 GitCode,发现更多精彩内容 >>
提交
414b5305
编写于
7月 19, 2021
作者:
N
niuyazhe
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
polish(nyz): fix repeat eval at beginning
上级
f962ef01
变更
4
显示空白变更内容
内联
并排
Showing
4 changed file
with
19 addition
and
14 deletion
+19
-14
ding/worker/collector/base_serial_evaluator.py
ding/worker/collector/base_serial_evaluator.py
+2
-0
dizoo/classic_control/bitflip/entry/bitflip_dqn_main.py
dizoo/classic_control/bitflip/entry/bitflip_dqn_main.py
+10
-9
dizoo/classic_control/cartpole/entry/cartpole_dqn_main.py
dizoo/classic_control/cartpole/entry/cartpole_dqn_main.py
+3
-2
dizoo/classic_control/pendulum/entry/pendulum_td3_main.py
dizoo/classic_control/pendulum/entry/pendulum_td3_main.py
+4
-3
未找到文件。
ding/worker/collector/base_serial_evaluator.py
浏览文件 @
414b5305
...
...
@@ -155,6 +155,8 @@ class BaseSerialEvaluator(object):
Determine whether you need to start the evaluation mode, if the number of training has reached
\
the maximum number of times to start the evaluator, return True
"""
if
train_iter
==
self
.
_last_eval_iter
:
return
False
if
(
train_iter
-
self
.
_last_eval_iter
)
<
self
.
_cfg
.
eval_freq
and
train_iter
!=
0
:
return
False
self
.
_last_eval_iter
=
train_iter
...
...
dizoo/classic_control/bitflip/entry/bitflip_dqn_main.py
浏览文件 @
414b5305
...
...
@@ -83,7 +83,8 @@ def main(cfg, seed=0, max_iterations=int(1e8)):
else
:
sample_size
=
learner
.
policy
.
get_attribute
(
'batch_size'
)
train_episode
=
replay_buffer
.
sample
(
sample_size
,
learner
.
train_iter
)
if
train_episode
is
not
None
:
if
train_episode
is
None
:
break
train_data
=
[]
if
her_cfg
is
not
None
:
her_episodes
=
[]
...
...
dizoo/classic_control/cartpole/entry/cartpole_dqn_main.py
浏览文件 @
414b5305
...
...
@@ -71,7 +71,8 @@ def main(cfg, seed=0):
# Training
for
i
in
range
(
cfg
.
policy
.
learn
.
update_per_collect
):
train_data
=
replay_buffer
.
sample
(
learner
.
policy
.
get_attribute
(
'batch_size'
),
learner
.
train_iter
)
if
train_data
is
not
None
:
if
train_data
is
None
:
break
learner
.
train
(
train_data
,
collector
.
envstep
)
...
...
dizoo/classic_control/pendulum/entry/pendulum_td3_main.py
浏览文件 @
414b5305
...
...
@@ -63,10 +63,11 @@ def main(cfg, seed=0):
# Collect data from environments
new_data
=
collector
.
collect
(
train_iter
=
learner
.
train_iter
)
replay_buffer
.
push
(
new_data
,
cur_collector_envstep
=
collector
.
envstep
)
# Tr
ia
n
# Tr
ai
n
for
i
in
range
(
cfg
.
policy
.
learn
.
update_per_collect
):
train_data
=
replay_buffer
.
sample
(
learner
.
policy
.
get_attribute
(
'batch_size'
),
learner
.
train_iter
)
if
train_data
is
not
None
:
if
train_data
is
None
:
break
learner
.
train
(
train_data
,
collector
.
envstep
)
...
...
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录