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
bae419ca
Commit
bae419ca
authored
Apr 23, 2019
by
Yuxin Wu
Browse files
Options
Browse Files
Download
Email Patches
Plain Diff
[MaskRCNN] move model architectures to generalized_rcnn.py (#1163)
parent
8ce73803
Changes
9
Expand all
Hide whitespace changes
Inline
Side-by-side
Showing
9 changed files
with
320 additions
and
302 deletions
+320
-302
examples/DeepQNetwork/README.md
examples/DeepQNetwork/README.md
+5
-3
examples/FasterRCNN/NOTES.md
examples/FasterRCNN/NOTES.md
+3
-2
examples/FasterRCNN/backbone.py
examples/FasterRCNN/backbone.py
+1
-1
examples/FasterRCNN/config.py
examples/FasterRCNN/config.py
+1
-1
examples/FasterRCNN/generalized_rcnn.py
examples/FasterRCNN/generalized_rcnn.py
+304
-0
examples/FasterRCNN/model_fpn.py
examples/FasterRCNN/model_fpn.py
+1
-1
examples/FasterRCNN/model_frcnn.py
examples/FasterRCNN/model_frcnn.py
+1
-1
examples/FasterRCNN/model_mrcnn.py
examples/FasterRCNN/model_mrcnn.py
+1
-1
examples/FasterRCNN/train.py
examples/FasterRCNN/train.py
+3
-292
No files found.
examples/DeepQNetwork/README.md
View file @
bae419ca
...
...
@@ -20,9 +20,11 @@ Claimed performance in the paper can be reproduced, on several games I've tested

On one GTX 1080Ti, the ALE version took __~2 hours__ of training to reach 21 (maximum) score on
Pong, __~10 hours__ of training to reach 400 score on Breakout.
It runs at 80 batches (~5.1k trained frames, 320 seen frames, 1.3k game frames) per second on GTX 1080Ti.
On one GTX 1080Ti,
the ALE version took
__~2 hours__
of training to reach 21 (maximum) score on Pong,
__~10 hours__
of training to reach 400 score on Breakout.
It runs at 100 batches (6.4k trained frames, 400 seen frames, 1.6k game frames) per second on GTX 1080Ti.
This is likely the fastest open source TF implementation of DQN.
## How to use
...
...
examples/FasterRCNN/NOTES.md
View file @
bae419ca
...
...
@@ -4,9 +4,10 @@ This is a minimal implementation that simply contains these files:
+
dataset.py: load and evaluate COCO dataset
+
data.py: prepare data for training & inference
+
common.py: common data preparation utilities
+
ba
semodel
.py: implement backbones
+
ba
ckbone
.py: implement backbones
+
model_box.py: implement box-related symbolic functions
+
model_{fpn,rpn,frcnn,mrcnn,cascade}.py: implement FPN,RPN,Fast-/Mask-/Cascade-RCNN models.
+
generalized_rcnn.py: implement variants of generalized R-CNN architecture
+
model_{fpn,rpn,frcnn,mrcnn,cascade}.py: implement FPN,RPN,Fast/Mask/Cascade R-CNN models.
+
train.py: main entry script
+
utils/: third-party helper functions
+
eval.py: evaluation utilities
...
...
examples/FasterRCNN/ba
semodel
.py
→
examples/FasterRCNN/ba
ckbone
.py
View file @
bae419ca
# -*- coding: utf-8 -*-
# File: ba
semodel
.py
# File: ba
ckbone
.py
import
numpy
as
np
from
contextlib
import
ExitStack
,
contextmanager
...
...
examples/FasterRCNN/config.py
View file @
bae419ca
...
...
@@ -95,7 +95,7 @@ _C.DATA.CLASS_NAMES = [] # NUM_CLASS (NUM_CATEGORY+1) strings, the first is "BG
_C
.
DATA
.
ABSOLUTE_COORD
=
True
_C
.
DATA
.
NUM_WORKERS
=
5
# number of data loading workers
# ba
semodel
----------------------
# ba
ckbone
----------------------
_C
.
BACKBONE
.
WEIGHTS
=
''
# /path/to/weights.npz
_C
.
BACKBONE
.
RESNET_NUM_BLOCKS
=
[
3
,
4
,
6
,
3
]
# for resnet50
# RESNET_NUM_BLOCKS = [3, 4, 23, 3] # for resnet101
...
...
examples/FasterRCNN/generalized_rcnn.py
0 → 100644
View file @
bae419ca
This diff is collapsed.
Click to expand it.
examples/FasterRCNN/model_fpn.py
View file @
bae419ca
...
...
@@ -10,7 +10,7 @@ from tensorpack.tfutils.scope_utils import under_name_scope
from
tensorpack.tfutils.summary
import
add_moving_summary
from
tensorpack.tfutils.tower
import
get_current_tower_context
from
ba
semodel
import
GroupNorm
from
ba
ckbone
import
GroupNorm
from
config
import
config
as
cfg
from
model_box
import
roi_align
from
model_rpn
import
generate_rpn_proposals
,
rpn_losses
...
...
examples/FasterRCNN/model_frcnn.py
View file @
bae419ca
...
...
@@ -10,7 +10,7 @@ from tensorpack.tfutils.scope_utils import under_name_scope
from
tensorpack.tfutils.summary
import
add_moving_summary
from
tensorpack.utils.argtools
import
memoized_method
from
ba
semodel
import
GroupNorm
from
ba
ckbone
import
GroupNorm
from
config
import
config
as
cfg
from
model_box
import
decode_bbox_target
,
encode_bbox_target
from
utils.box_ops
import
pairwise_iou
...
...
examples/FasterRCNN/model_mrcnn.py
View file @
bae419ca
...
...
@@ -8,7 +8,7 @@ from tensorpack.tfutils.common import get_tf_version_tuple
from
tensorpack.tfutils.scope_utils
import
under_name_scope
from
tensorpack.tfutils.summary
import
add_moving_summary
from
ba
semodel
import
GroupNorm
from
ba
ckbone
import
GroupNorm
from
config
import
config
as
cfg
...
...
examples/FasterRCNN/train.py
View file @
bae419ca
This diff is collapsed.
Click to expand it.
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