Skip to content
Projects
Groups
Snippets
Help
Loading...
Help
Support
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in
Toggle navigation
S
seminar-breakout
Project overview
Project overview
Details
Activity
Releases
Repository
Repository
Files
Commits
Branches
Tags
Contributors
Graph
Compare
Issues
0
Issues
0
List
Boards
Labels
Milestones
Merge Requests
0
Merge Requests
0
CI / CD
CI / CD
Pipelines
Jobs
Schedules
Analytics
Analytics
CI / CD
Repository
Value Stream
Wiki
Wiki
Members
Members
Collapse sidebar
Close sidebar
Activity
Graph
Create a new issue
Jobs
Commits
Issue Boards
Open sidebar
Shashank Suhas
seminar-breakout
Commits
776c0dbb
Commit
776c0dbb
authored
Aug 24, 2016
by
Yuxin Wu
Browse files
Options
Browse Files
Download
Email Patches
Plain Diff
bug fix in DQN
parent
266ac578
Changes
3
Hide whitespace changes
Inline
Side-by-side
Showing
3 changed files
with
6 additions
and
4 deletions
+6
-4
examples/Atari2600/DQN.py
examples/Atari2600/DQN.py
+1
-1
examples/OpenAIGym/README.md
examples/OpenAIGym/README.md
+4
-3
examples/cifar-convnet.py
examples/cifar-convnet.py
+1
-0
No files found.
examples/Atari2600/DQN.py
View file @
776c0dbb
...
...
@@ -115,7 +115,7 @@ class Model(ModelDesc):
target
=
reward
+
(
1.0
-
tf
.
cast
(
isOver
,
tf
.
float32
))
*
GAMMA
*
tf
.
stop_gradient
(
best_v
)
cost
=
symbf
.
clipped_l2
_loss
(
target
-
pred_action_value
)
cost
=
symbf
.
huber
_loss
(
target
-
pred_action_value
)
summary
.
add_param_summary
([(
'conv.*/W'
,
[
'histogram'
,
'rms'
]),
(
'fc.*/W'
,
[
'histogram'
,
'rms'
])
])
# monitor all W
self
.
cost
=
tf
.
reduce_mean
(
cost
,
name
=
'cost'
)
...
...
examples/OpenAIGym/README.md
View file @
776c0dbb
...
...
@@ -14,9 +14,10 @@ Models are available for the following gym atari environments (click links for v
+
[
Asterix-v0
](
https://gym.openai.com/evaluations/eval_mees2c58QfKm5GspCjRfCA
)
+
[
Asteroids-v0
](
https://gym.openai.com/evaluations/eval_8eHKsRL4RzuZEq9AOLZA
)
+
[
Atlantis-v0
](
https://gym.openai.com/evaluations/eval_Z1B3d7A1QCaQk1HpO1Rg
)
+
[
BattleZone-v0
](
https://gym.openai.com/evaluations/eval_SoLit2bR1qmFoC0AsJF6Q
)
+
[
BankHeist-v0
](
https://gym.openai.com/evaluations/eval_hifoaxFTIuLlPd38BjnOw
)
+
[
BattleZone-v0
](
https://gym.openai.com/evaluations/eval_SoLit2bR1qmFoC0AsJF6Q
)
+
[
BeamRider-v0
](
https://gym.openai.com/evaluations/eval_KuOYumrjQjixwL0spG0iCA
)
+
[
Berzerk-v0
](
https://gym.openai.com/evaluations/eval_Yri0XQbwRy62NzWILdn5IA
)
+
[
Breakout-v0
](
https://gym.openai.com/evaluations/eval_L55gczPrQJamMGihq9tzA
)
+
[
Carnival-v0
](
https://gym.openai.com/evaluations/eval_xJSOlo2lSWaH1wHEOX5vw
)
+
[
ChopperCommand-v0
](
https://gym.openai.com/evaluations/eval_tYVKyh7wQieRIKgEvVaCuw
)
...
...
@@ -31,9 +32,9 @@ Models are available for the following gym atari environments (click links for v
+
[
Krull-v0
](
https://gym.openai.com/evaluations/eval_dfOS2WzhTh6sn1FuPS9HA
)
+
[
KungFuMaster-v0
](
https://gym.openai.com/evaluations/eval_vNWDShYTRC0MhfIybeUYg
)
+
[
MsPacman-v0
](
https://gym.openai.com/evaluations/eval_kpL9bSsS4GXsYb9HuEfew
)
+
[
Pooyan-v0
](
https://gym.openai.com/evaluations/eval_UXFVI34MSAuNTtjZcK8N0A
)
+
[
Pong-v0
](
https://gym.openai.com/evaluations/eval_8L7SV59nSW6GGbbP3N4G6w
)
+
[
Phoenix-v0
](
https://gym.openai.com/evaluations/eval_uzUruiB3RRKUMvJIxvEzYA
)
+
[
Pong-v0
](
https://gym.openai.com/evaluations/eval_8L7SV59nSW6GGbbP3N4G6w
)
+
[
Pooyan-v0
](
https://gym.openai.com/evaluations/eval_UXFVI34MSAuNTtjZcK8N0A
)
+
[
Qbert-v0
](
https://gym.openai.com/evaluations/eval_wekCJkrWQm9NrOUzltXg
)
+
[
Riverraid-v0
](
https://gym.openai.com/evaluations/eval_OU4x3DkTfm4uaXy6CIaXg
)
+
[
RoadRunner-v0
](
https://gym.openai.com/evaluations/eval_wINKQTwxT9ipydHOXBhg
)
...
...
examples/cifar-convnet.py
View file @
776c0dbb
...
...
@@ -155,3 +155,4 @@ if __name__ == '__main__':
if
args
.
gpu
:
config
.
nr_tower
=
len
(
args
.
gpu
.
split
(
','
))
QueueInputTrainer
(
config
)
.
train
()
#AsyncMultiGPUTrainer(config).train()
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment