Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Fix rcnn target generator #1508

Merged
merged 4 commits into from
Nov 10, 2020
Merged

Fix rcnn target generator #1508

merged 4 commits into from
Nov 10, 2020

Conversation

zhreshold
Copy link
Member

Fix the issue of intialization error when target generator is not used at all.

@mli
Copy link
Member

mli commented Nov 10, 2020

Job PR-1508-4 is done.
Docs are uploaded to http://gluon-vision-staging.s3-website-us-west-2.amazonaws.com/PR-1508/4/index.html
Code coverage of this PR: pr.svg vs. Master: master.svg

@zhreshold zhreshold merged commit 75aae80 into dmlc:master Nov 10, 2020
@zhreshold zhreshold deleted the fix-rcnn-encoder-mean-std branch November 10, 2020 06:27
@chinakook
Copy link
Member

@zhreshold
This PR may cause the scripts/detection/faster_rcnn/train_faster_rcnn.py script to get RuntimeError.
RuntimeError: Parameter 'fasterrcnn0_normalizedperclassboxcenterencoder0_means' has not been initialized. Note that you should initialize parameters and create Trainer with Block.collect_params() instead of Block.params because the later does not include Parameters of nested child Block

yinweisu added a commit to yinweisu/gluon-cv that referenced this pull request Nov 18, 2020
@yinweisu @zhreshold
@bryanyzhu @Arthurlxy @ECHO960 @coocoo90 @BebDong @KuangHaofei
[WIP] Github Actions (#1)
4e67616 
* incorporate autodatasets (dmlc#1496)

* Add torch clarification (dmlc#1495)

* Add torch clarification

* fix

* Fix auto detectors (dmlc#1497)

* fix yolo predictor

* fix predict

* fix config (dmlc#1498)

* Added support for AWS Batch. Added support for docker (dmlc#1474)

* Added support for AWS Batch. Added support for docker

* Fixed style. Removed code in commet. Updated README to include boto3 usage

* Renamed template file. Removed gluon aws id

* fix readme

* fix

* fix imports (dmlc#1499)

* fix imports

* fix

* fix image classification

* fix

* fix width height

* fix

* fix batch size

* fix

* fix

* none to empty string (dmlc#1502)

* [WIP] Tinycoco (dmlc#1501)

* Add minicoco

* update jenkins for minicoco

* fix

* renamed mini to tiny

* fix

* fix

* fix, add VOCDetectionTiny

* fix

* fix env

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* test

* test

* test

* clean up

Co-authored-by: Joshua Z. Zhang <[email protected]>

* Fix rcnn target generator (dmlc#1508)

* fix not used rcnn target generator

* fix lint

* fix

* fix

* add get flops (dmlc#1509)

* warmup scheduler for video torch (dmlc#1510)

1. refine warmup logic, now using cfg.CONFIG.TRAIN.USE_WARMUP to control open warmup or not.
2. fix bug in gluoncv/torch/utils/lr_policy.py
3. change training configs
4. change ddp_train_pytorch and ddp_train_shortonly_pytorch, This is tested on ec2 machines

* update torchvideo model zoo (dmlc#1513)

* add ir-csn-152 into torchvideo model zoo (dmlc#1515)

* Revise danet.py (dmlc#1507)

The dropout layer should be placed before the classification layer.

* icnet missing background class (dmlc#1518)

* Add CSN model to torch video model zoo (dmlc#1517)

* add ircsn

* update model zoo

* fix lint

* Improve auto tasks (dmlc#1523)

* use in-memory pickle instead of disk file

* add feature extractor for image classification

* add tests

* fix

* fix lint

* more unittests

* fix

* fix

* Added github action and workflow for sanity check

* Removed container and actions.

* Added unit test

* Added build docs

* Fix

* Fix

* Fix

* Fix

* Test

* test

* Update unit test

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* subclass coco

* fix

* fix

* fix

* fix

* rebase conflict

* fix rebase

* fix

* fix

* add aws authentication

* add aws authentication

* test

* test

* test

* test

* test

* fix log

* test

* test

* test

* test

* test

* test

* fix

* rebase

* add tiny motorbike

* fix

* model zoo

* test

* fix docker

* parallel jobs

* parallel jobs

* fix

* add torch

* add torch

* fix

* fix

* fix

* full test

* full test

* test build docs

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* test branch

* test branch

* fix

* test

* test

* add comment

Co-authored-by: Joshua Z. Zhang <[email protected]>
Co-authored-by: Yi Zhu <[email protected]>
Co-authored-by: Xinyu Li <[email protected]>
Co-authored-by: Chunhui Liu <[email protected]>
Co-authored-by: YANYI ZHANG <[email protected]>
Co-authored-by: BebDong <[email protected]>
Co-authored-by: Kuang Haofei <[email protected]>
yinweisu added a commit to yinweisu/gluon-cv that referenced this pull request Nov 24, 2020
@yinweisu @zhreshold
@bryanyzhu @Arthurlxy @ECHO960 @coocoo90 @BebDong @KuangHaofei
[WIP] Github Actions (#1)
a94aba8 
* incorporate autodatasets (dmlc#1496)

* Add torch clarification (dmlc#1495)

* Add torch clarification

* fix

* Fix auto detectors (dmlc#1497)

* fix yolo predictor

* fix predict

* fix config (dmlc#1498)

* Added support for AWS Batch. Added support for docker (dmlc#1474)

* Added support for AWS Batch. Added support for docker

* Fixed style. Removed code in commet. Updated README to include boto3 usage

* Renamed template file. Removed gluon aws id

* fix readme

* fix

* fix imports (dmlc#1499)

* fix imports

* fix

* fix image classification

* fix

* fix width height

* fix

* fix batch size

* fix

* fix

* none to empty string (dmlc#1502)

* [WIP] Tinycoco (dmlc#1501)

* Add minicoco

* update jenkins for minicoco

* fix

* renamed mini to tiny

* fix

* fix

* fix, add VOCDetectionTiny

* fix

* fix env

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* test

* test

* test

* clean up

Co-authored-by: Joshua Z. Zhang <[email protected]>

* Fix rcnn target generator (dmlc#1508)

* fix not used rcnn target generator

* fix lint

* fix

* fix

* add get flops (dmlc#1509)

* warmup scheduler for video torch (dmlc#1510)

1. refine warmup logic, now using cfg.CONFIG.TRAIN.USE_WARMUP to control open warmup or not.
2. fix bug in gluoncv/torch/utils/lr_policy.py
3. change training configs
4. change ddp_train_pytorch and ddp_train_shortonly_pytorch, This is tested on ec2 machines

* update torchvideo model zoo (dmlc#1513)

* add ir-csn-152 into torchvideo model zoo (dmlc#1515)

* Revise danet.py (dmlc#1507)

The dropout layer should be placed before the classification layer.

* icnet missing background class (dmlc#1518)

* Add CSN model to torch video model zoo (dmlc#1517)

* add ircsn

* update model zoo

* fix lint

* Improve auto tasks (dmlc#1523)

* use in-memory pickle instead of disk file

* add feature extractor for image classification

* add tests

* fix

* fix lint

* more unittests

* fix

* fix

* Added github action and workflow for sanity check

* Removed container and actions.

* Added unit test

* Added build docs

* Fix

* Fix

* Fix

* Fix

* Test

* test

* Update unit test

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* subclass coco

* fix

* fix

* fix

* fix

* rebase conflict

* fix rebase

* fix

* fix

* add aws authentication

* add aws authentication

* test

* test

* test

* test

* test

* fix log

* test

* test

* test

* test

* test

* test

* fix

* rebase

* add tiny motorbike

* fix

* model zoo

* test

* fix docker

* parallel jobs

* parallel jobs

* fix

* add torch

* add torch

* fix

* fix

* fix

* full test

* full test

* test build docs

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* test branch

* test branch

* fix

* test

* test

* add comment

Co-authored-by: Joshua Z. Zhang <[email protected]>
Co-authored-by: Yi Zhu <[email protected]>
Co-authored-by: Xinyu Li <[email protected]>
Co-authored-by: Chunhui Liu <[email protected]>
Co-authored-by: YANYI ZHANG <[email protected]>
Co-authored-by: BebDong <[email protected]>
Co-authored-by: Kuang Haofei <[email protected]>
zhreshold added a commit that referenced this pull request Dec 2, 2020
@yinweisu @zhreshold
@bryanyzhu @Arthurlxy @ECHO960 @coocoo90 @BebDong @KuangHaofei
Github Actions for CI (#1541)
386be93 
* [WIP] Github Actions (#1)

* incorporate autodatasets (#1496)

* Add torch clarification (#1495)

* Add torch clarification

* fix

* Fix auto detectors (#1497)

* fix yolo predictor

* fix predict

* fix config (#1498)

* Added support for AWS Batch. Added support for docker (#1474)

* Added support for AWS Batch. Added support for docker

* Fixed style. Removed code in commet. Updated README to include boto3 usage

* Renamed template file. Removed gluon aws id

* fix readme

* fix

* fix imports (#1499)

* fix imports

* fix

* fix image classification

* fix

* fix width height

* fix

* fix batch size

* fix

* fix

* none to empty string (#1502)

* [WIP] Tinycoco (#1501)

* Add minicoco

* update jenkins for minicoco

* fix

* renamed mini to tiny

* fix

* fix

* fix, add VOCDetectionTiny

* fix

* fix env

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* test

* test

* test

* clean up

Co-authored-by: Joshua Z. Zhang <[email protected]>

* Fix rcnn target generator (#1508)

* fix not used rcnn target generator

* fix lint

* fix

* fix

* add get flops (#1509)

* warmup scheduler for video torch (#1510)

1. refine warmup logic, now using cfg.CONFIG.TRAIN.USE_WARMUP to control open warmup or not.
2. fix bug in gluoncv/torch/utils/lr_policy.py
3. change training configs
4. change ddp_train_pytorch and ddp_train_shortonly_pytorch, This is tested on ec2 machines

* update torchvideo model zoo (#1513)

* add ir-csn-152 into torchvideo model zoo (#1515)

* Revise danet.py (#1507)

The dropout layer should be placed before the classification layer.

* icnet missing background class (#1518)

* Add CSN model to torch video model zoo (#1517)

* add ircsn

* update model zoo

* fix lint

* Improve auto tasks (#1523)

* use in-memory pickle instead of disk file

* add feature extractor for image classification

* add tests

* fix

* fix lint

* more unittests

* fix

* fix

* Added github action and workflow for sanity check

* Removed container and actions.

* Added unit test

* Added build docs

* Fix

* Fix

* Fix

* Fix

* Test

* test

* Update unit test

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* subclass coco

* fix

* fix

* fix

* fix

* rebase conflict

* fix rebase

* fix

* fix

* add aws authentication

* add aws authentication

* test

* test

* test

* test

* test

* fix log

* test

* test

* test

* test

* test

* test

* fix

* rebase

* add tiny motorbike

* fix

* model zoo

* test

* fix docker

* parallel jobs

* parallel jobs

* fix

* add torch

* add torch

* fix

* fix

* fix

* full test

* full test

* test build docs

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* test branch

* test branch

* fix

* test

* test

* add comment

Co-authored-by: Joshua Z. Zhang <[email protected]>
Co-authored-by: Yi Zhu <[email protected]>
Co-authored-by: Xinyu Li <[email protected]>
Co-authored-by: Chunhui Liu <[email protected]>
Co-authored-by: YANYI ZHANG <[email protected]>
Co-authored-by: BebDong <[email protected]>
Co-authored-by: Kuang Haofei <[email protected]>

* [WIP] Test PR (#3)

* Added github action and workflow for sanity check

* Removed container and actions.

* Added unit test

* Added build docs

* Fix

* Fix

* Fix

* Fix

* Test

* test

* Update unit test

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* subclass coco

* fix

* fix

* fix

* fix

* rebase conflict

* fix rebase

* fix

* fix

* add aws authentication

* add aws authentication

* test

* test

* test

* test

* test

* fix log

* test

* test

* test

* test

* test

* test

* fix

* rebase

* add tiny motorbike

* fix

* model zoo

* test

* fix docker

* parallel jobs

* parallel jobs

* fix

* add torch

* add torch

* fix

* fix

* fix

* full test

* full test

* test build docs

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* test branch

* test branch

* fix

* test

* test

* add comment

* test

* full test

* full test

* full test

* test (#5)

* test

* fix

* change to 12x

* test comments

* change to pr_target

* [WIP] Full Test (#6)

* full test

* test model zoo

* test model zoo

* full test

* full test

* add auto

* add gpu_test.sh

* test efs modelzoo

* test efs modelzoo

* test efs modelzoo

* test without auto

* test repo name

* test repo name

* test repo name

* test repo name

* test sharemem

* full test (#8)

* [WIP] Github Actions (#1)

* incorporate autodatasets (#1496)

* Add torch clarification (#1495)

* Add torch clarification

* fix

* Fix auto detectors (#1497)

* fix yolo predictor

* fix predict

* fix config (#1498)

* Added support for AWS Batch. Added support for docker (#1474)

* Added support for AWS Batch. Added support for docker

* Fixed style. Removed code in commet. Updated README to include boto3 usage

* Renamed template file. Removed gluon aws id

* fix readme

* fix

* fix imports (#1499)

* fix imports

* fix

* fix image classification

* fix

* fix width height

* fix

* fix batch size

* fix

* fix

* none to empty string (#1502)

* [WIP] Tinycoco (#1501)

* Add minicoco

* update jenkins for minicoco

* fix

* renamed mini to tiny

* fix

* fix

* fix, add VOCDetectionTiny

* fix

* fix env

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* test

* test

* test

* clean up

Co-authored-by: Joshua Z. Zhang <[email protected]>

* Fix rcnn target generator (#1508)

* fix not used rcnn target generator

* fix lint

* fix

* fix

* add get flops (#1509)

* warmup scheduler for video torch (#1510)

1. refine warmup logic, now using cfg.CONFIG.TRAIN.USE_WARMUP to control open warmup or not.
2. fix bug in gluoncv/torch/utils/lr_policy.py
3. change training configs
4. change ddp_train_pytorch and ddp_train_shortonly_pytorch, This is tested on ec2 machines

* update torchvideo model zoo (#1513)

* add ir-csn-152 into torchvideo model zoo (#1515)

* Revise danet.py (#1507)

The dropout layer should be placed before the classification layer.

* icnet missing background class (#1518)

* Add CSN model to torch video model zoo (#1517)

* add ircsn

* update model zoo

* fix lint

* Improve auto tasks (#1523)

* use in-memory pickle instead of disk file

* add feature extractor for image classification

* add tests

* fix

* fix lint

* more unittests

* fix

* fix

* Added github action and workflow for sanity check

* Removed container and actions.

* Added unit test

* Added build docs

* Fix

* Fix

* Fix

* Fix

* Test

* test

* Update unit test

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* subclass coco

* fix

* fix

* fix

* fix

* rebase conflict

* fix rebase

* fix

* fix

* add aws authentication

* add aws authentication

* test

* test

* test

* test

* test

* fix log

* test

* test

* test

* test

* test

* test

* fix

* rebase

* add tiny motorbike

* fix

* model zoo

* test

* fix docker

* parallel jobs

* parallel jobs

* fix

* add torch

* add torch

* fix

* fix

* fix

* full test

* full test

* test build docs

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* test branch

* test branch

* fix

* test

* test

* add comment

Co-authored-by: Joshua Z. Zhang <[email protected]>
Co-authored-by: Yi Zhu <[email protected]>
Co-authored-by: Xinyu Li <[email protected]>
Co-authored-by: Chunhui Liu <[email protected]>
Co-authored-by: YANYI ZHANG <[email protected]>
Co-authored-by: BebDong <[email protected]>
Co-authored-by: Kuang Haofei <[email protected]>

* [WIP] Test PR (#3)

* Added github action and workflow for sanity check

* Removed container and actions.

* Added unit test

* Added build docs

* Fix

* Fix

* Fix

* Fix

* Test

* test

* Update unit test

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* subclass coco

* fix

* fix

* fix

* fix

* rebase conflict

* fix rebase

* fix

* fix

* add aws authentication

* add aws authentication

* test

* test

* test

* test

* test

* fix log

* test

* test

* test

* test

* test

* test

* fix

* rebase

* add tiny motorbike

* fix

* model zoo

* test

* fix docker

* parallel jobs

* parallel jobs

* fix

* add torch

* add torch

* fix

* fix

* fix

* full test

* full test

* test build docs

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* test branch

* test branch

* fix

* test

* test

* add comment

* test

* full test

* full test

* full test

* test (#5)

* test

* fix

* change to 12x

* test comments

* change to pr_target

* [WIP] Full Test (#6)

* full test

* test model zoo

* test model zoo

* full test

* full test

* add auto

* add gpu_test.sh

* test efs modelzoo

* test efs modelzoo

* test efs modelzoo

* test without auto

* test repo name

* test repo name

* test repo name

* test repo name

* test sharemem

* test pr only on yinweisu

* test pr only on yinweisu

* update repo name

* test pr only on yinweisu (#9)

* full test on pr only yinweisu (#10)

* ready to pr

* fix

* change doc env name

* add torch to env

* add yacs to env

* fix path

Co-authored-by: Joshua Z. Zhang <[email protected]>
Co-authored-by: Yi Zhu <[email protected]>
Co-authored-by: Xinyu Li <[email protected]>
Co-authored-by: Chunhui Liu <[email protected]>
Co-authored-by: YANYI ZHANG <[email protected]>
Co-authored-by: BebDong <[email protected]>
Co-authored-by: Kuang Haofei <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@zhreshold @mli @chinakook