profile
viewpoint
If you are wondering where the data of this site comes from, please visit https://api.github.com/users/reactivetype/events. GitMemory does not store any data, but only uses NGINX to cache data for a period of time. The idea behind GitMemory is simply to give users a better reading experience.
Rindra Ramamonjison reactivetype

reactivetype/elm-dragNdrop 4

Elm Drag and Drop examples

reactivetype/DivideMix 2

Code for paper: DivideMix: Learning with Noisy Labels as Semi-supervised Learning

reactivetype/AB_distillation 0

Knowledge Transfer via Distillation of Activation Boundaries Formed by Hidden Neurons (AAAI 2019)

reactivetype/awesome-courses 0

:books: List of awesome university courses for learning Computer Science!

reactivetype/awesome-emdl 0

Embedded and mobile deep learning research resources

reactivetype/awesome-nlp 0

:book: A curated list of resources dedicated to Natural Language Processing (NLP)

reactivetype/bottleneck-transformer-pytorch 0

Implementation of Bottleneck Transformer in Pytorch

reactivetype/brocolli 0

pytorch 2 caffe

reactivetype/CenterNet 0

Codes for our paper "CenterNet: Keypoint Triplets for Object Detection" .

issue openedultralytics/yolov5

How to train on 128*4500 pictures

โ”Question

I donโ€™t want it to fill 128 to 4500, it consumes resources too much

Additional context

created time in an hour

PR closed PyTorchLightning/pytorch-lightning

[WIP] Allow loggers to be restarted on HPC resume won't fix

What does this PR do?

Fixes #6205

This problem is not restricted to MLFlow as discussed in the issue though, it is a problem with pretty much every logger.

This PR fixes the problem of loggers not resuming when an experiment running on HPC (i.e., slurm) times out and is requeued. The same experiment is requeued and resumes running, but the loggers were making a new experiment entry on the UI. I fixed this by saving the logger in the HPC checkpoint and then moving the HPC resume call earlier in the trainer code.

This PR is heavily a WIP and I have marked it as such, there's a few small things missing but mainly I wanted to get the ball rolling on some discussion around it. I'm using this code to train real models and it works fine for me but I would love someone who knows all the implications of moving the HPC loading code around to have a look at this. I tried to start this discussion on #6205 but I didnt get any replies despite repeatedly asking.

Before submitting

  • [x] Was this discussed/approved via a GitHub issue? (not for typos and docs)
  • [x] Did you read the contributor guideline, Pull Request section?
  • [x] Did you make sure your PR does only one thing, instead of bundling different changes together?
  • [x] Did you make sure to update the documentation with your changes? (if necessary)
  • [ ] Did you write any new necessary tests? (not for typos and docs)
  • [ ] Did you verify new and existing tests pass locally with your changes?
  • [ ] Did you update the CHANGELOG? (not for typos, docs, test updates, or internal minor changes/refactorings)

<!-- For CHANGELOG separate each item in the unreleased section by a blank line to reduce collisions -->

PR review

Anyone in the community is free to review the PR once the tests have passed. Before you start reviewing make sure you have read Review guidelines. In short, see the following bullet-list:

  • [ ] Is this pull request ready for review? (if not, please submit in draft mode)
  • [ ] Check that all items from Before submitting are resolved
  • [ ] Make sure the title is self-explanatory and the description concisely explains the PR
  • [ ] Add labels and milestones (and optionally projects) to the PR so it can be classified

Did you have fun?

Make sure you had fun coding ๐Ÿ™ƒ

+55 -3

10 comments

4 changed files

Queuecumber

pr closed time in 2 hours

pull request commentPyTorchLightning/pytorch-lightning

[WIP] Allow loggers to be restarted on HPC resume

This pull request is going to be closed. Please feel free to reopen it create a new from the actual master.

Queuecumber

comment created time in 2 hours

pull request commentPyTorchLightning/pytorch-lightning

Changes to load model state in checkpoint connector in case of multiple workers

Codecov Report

Merging #8044 (20e429a) into master (c7eaf76) will decrease coverage by 47%. The diff coverage is 19%.

@@           Coverage Diff            @@
##           master   #8044     +/-   ##
========================================
- Coverage      92%     45%    -47%     
========================================
  Files         210     210             
  Lines       13579   13593     +14     
========================================
- Hits        12452    6111   -6341     
- Misses       1127    7482   +6355     
mleshen

comment created time in 4 hours

pull request commentPyTorchLightning/pytorch-lightning

Changes to load model state in checkpoint connector in case of multiple workers

Hello @mleshen! Thanks for opening this PR.

Line 167:121: E501 line too long (130 > 120 characters) Line 175:5: E303 too many blank lines (2)

Do see the Hitchhiker's guide to code style

mleshen

comment created time in 4 hours

PR opened PyTorchLightning/pytorch-lightning

Changes to load model state in checkpoint connector in case of multiple workers

What does this PR do?

Proposed solution to #<#8043>

Before submitting

  • [x] Was this discussed/approved via a GitHub issue? (not for typos and docs)
  • [x] Did you read the contributor guideline, Pull Request section?
  • [x] Did you make sure your PR does only one thing, instead of bundling different changes together?
  • [ ] Did you make sure to update the documentation with your changes? (if necessary)
  • [ ] Did you write any new necessary tests? (not for typos and docs)
  • [ ] Did you verify new and existing tests pass locally with your changes?
  • [ ] Did you update the CHANGELOG? (not for typos, docs, test updates, or internal minor changes/refactorings)

<!-- For CHANGELOG separate each item in the unreleased section by a blank line to reduce collisions -->

PR review

Anyone in the community is free to review the PR once the tests have passed. Before you start reviewing make sure you have read Review guidelines. In short, see the following bullet-list:

  • [ ] Is this pull request ready for review? (if not, please submit in draft mode)
  • [ ] Check that all items from Before submitting are resolved
  • [ ] Make sure the title is self-explanatory and the description concisely explains the PR
  • [ ] Add labels and milestones (and optionally projects) to the PR so it can be classified

Did you have fun?

Make sure you had fun coding ๐Ÿ™ƒ

+22 -0

0 comment

1 changed file

pr created time in 4 hours

issue openedPyTorchLightning/pytorch-lightning

OOM issues with loading large model checkpoints w/ FSDP after checkpoint refactor

๐Ÿ› Bug

In #7928 the trainer logic was modified to restore the model state from the checkpoint connector instead of from the training type plugin and restore_model_from_ckpt_path was split into three new modular APIs. For our use case we overrode restore_model_from_ckpt_path in the FSDP plugin to prevent CPU OOMs, and now that the functionality for restoring the model state has been offloaded to the checkpoint, we run into OOMs again.

In #7509 it was proposed to solve this problem on the level of trainer โ€” comment suggests offloading responsibility to training_type_plugin since this is not widely required outside of DDP and its derivatives, but restoring model state functionality no longer belongs to the plugin. Could we add some more memory-friendly logic to the checkpoint connector in case of multiple workers?

Please reproduce using the BoringModel

<!-- Please paste your BoringModel colab link here. -->

To Reproduce

Use following BoringModel and post here

<!-- If you could not reproduce using the BoringModel and still think there's a bug, please post here -->

Expected behavior

<!-- FILL IN -->

Environment

Note: Bugs with code are solved faster ! Colab Notebook should be made public !

You can get the script and run it with:

wget https://raw.githubusercontent.com/PyTorchLightning/pytorch-lightning/master/tests/collect_env_details.py
# For security purposes, please check the contents of collect_env_details.py before running it.
python collect_env_details.py
  • PyTorch Version (e.g., 1.0):
  • OS (e.g., Linux):
  • How you installed PyTorch (conda, pip, source):
  • Build command you used (if compiling from source):
  • Python version:
  • CUDA/cuDNN version:
  • GPU models and configuration:
  • Any other relevant information:

Additional context

<!-- Add any other context about the problem here. -->

created time in 4 hours

issue commentultralytics/yolov5

Getting Segmentation Fault Error(core dumped) after successfully installing the coco datasets

@glenn-jocher Why are you not responding on this issue? I am not sure but your response can help me move ahead from this stuck problem

jaskiratsingh2000

comment created time in 6 hours

issue closedultralytics/yolov5

CVPR 2021 - Streaming Perception Challenge

We've just launched the first โ€œStreaming Perception Challengeโ€ at the Workshop on Autonomous Driving (WAD) in conjunction with CVPR 2021. The challenge is hosted by a team from CMU & UIUC and includes two tracks for streaming object detection: detection-only and full-stack (detection + tracking + forecasting). This challenge is to foster research in the area of Streaming Perception, which has garnered a lot of research interest after the paper โ€œTowards Streaming Perceptionโ€ was published last year. It received the Best Paper Honorable Mention award at ECCV 2020. Unlike most existing challenges, algorithm latency will be scored together with accuracy in a coherent manner. More details can be found on the challenge website. The total prize pool is $2700.

Please consider attending! If you have any questions, please feel free to contact us on the email address given on the website.

For new participants, we recommend that you check out the project website: http://www.cs.cmu.edu/~mengtial/proj/streaming/

closed time in 6 hours

karthiksharma98

issue commentultralytics/yolov5

Criteria for selecting the best model?

๐Ÿ‘‹ Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 ๐Ÿš€ resources:

  • Wiki โ€“ https://github.com/ultralytics/yolov5/wiki
  • Tutorials โ€“ https://github.com/ultralytics/yolov5#tutorials
  • Docs โ€“ https://docs.ultralytics.com

Access additional Ultralytics โšก resources:

  • Ultralytics HUB โ€“ https://ultralytics.com
  • Vision API โ€“ https://ultralytics.com/yolov5
  • About Us โ€“ https://ultralytics.com/about
  • Join Our Team โ€“ https://ultralytics.com/work
  • Contact Us โ€“ https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 ๐Ÿš€ and Vision AI โญ!

dariogonle

comment created time in 6 hours

issue commentultralytics/yolov5

YOLOv5 mAP and Precision Calculation

๐Ÿ‘‹ Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 ๐Ÿš€ resources:

  • Wiki โ€“ https://github.com/ultralytics/yolov5/wiki
  • Tutorials โ€“ https://github.com/ultralytics/yolov5#tutorials
  • Docs โ€“ https://docs.ultralytics.com

Access additional Ultralytics โšก resources:

  • Ultralytics HUB โ€“ https://ultralytics.com
  • Vision API โ€“ https://ultralytics.com/yolov5
  • About Us โ€“ https://ultralytics.com/about
  • Join Our Team โ€“ https://ultralytics.com/work
  • Contact Us โ€“ https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 ๐Ÿš€ and Vision AI โญ!

edahall

comment created time in 6 hours

issue commentultralytics/yolov5

Classification accuracy low

๐Ÿ‘‹ Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 ๐Ÿš€ resources:

  • Wiki โ€“ https://github.com/ultralytics/yolov5/wiki
  • Tutorials โ€“ https://github.com/ultralytics/yolov5#tutorials
  • Docs โ€“ https://docs.ultralytics.com

Access additional Ultralytics โšก resources:

  • Ultralytics HUB โ€“ https://ultralytics.com
  • Vision API โ€“ https://ultralytics.com/yolov5
  • About Us โ€“ https://ultralytics.com/about
  • Join Our Team โ€“ https://ultralytics.com/work
  • Contact Us โ€“ https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 ๐Ÿš€ and Vision AI โญ!

XingZeng307

comment created time in 6 hours

issue commentultralytics/yolov5

cant test with test.py

๐Ÿ‘‹ Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 ๐Ÿš€ resources:

  • Wiki โ€“ https://github.com/ultralytics/yolov5/wiki
  • Tutorials โ€“ https://github.com/ultralytics/yolov5#tutorials
  • Docs โ€“ https://docs.ultralytics.com

Access additional Ultralytics โšก resources:

  • Ultralytics HUB โ€“ https://ultralytics.com
  • Vision API โ€“ https://ultralytics.com/yolov5
  • About Us โ€“ https://ultralytics.com/about
  • Join Our Team โ€“ https://ultralytics.com/work
  • Contact Us โ€“ https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 ๐Ÿš€ and Vision AI โญ!

alicera

comment created time in 6 hours

issue commentultralytics/yolov5

RuntimeError: The size of tensor a (255) must match the size of tensor b (18) at non-singleton dimension 0

๐Ÿ‘‹ Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 ๐Ÿš€ resources:

  • Wiki โ€“ https://github.com/ultralytics/yolov5/wiki
  • Tutorials โ€“ https://github.com/ultralytics/yolov5#tutorials
  • Docs โ€“ https://docs.ultralytics.com

Access additional Ultralytics โšก resources:

  • Ultralytics HUB โ€“ https://ultralytics.com
  • Vision API โ€“ https://ultralytics.com/yolov5
  • About Us โ€“ https://ultralytics.com/about
  • Join Our Team โ€“ https://ultralytics.com/work
  • Contact Us โ€“ https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 ๐Ÿš€ and Vision AI โญ!

Rianusr

comment created time in 6 hours

issue commentultralytics/yolov5

Model Ensembling?

๐Ÿ‘‹ Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 ๐Ÿš€ resources:

  • Wiki โ€“ https://github.com/ultralytics/yolov5/wiki
  • Tutorials โ€“ https://github.com/ultralytics/yolov5#tutorials
  • Docs โ€“ https://docs.ultralytics.com

Access additional Ultralytics โšก resources:

  • Ultralytics HUB โ€“ https://ultralytics.com
  • Vision API โ€“ https://ultralytics.com/yolov5
  • About Us โ€“ https://ultralytics.com/about
  • Join Our Team โ€“ https://ultralytics.com/work
  • Contact Us โ€“ https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 ๐Ÿš€ and Vision AI โญ!

dariogonle

comment created time in 6 hours

issue commentultralytics/yolov5

Two Erros if i train custom dataset in Jupypter Notebook tutorial.ipynb

๐Ÿ‘‹ Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 ๐Ÿš€ resources:

  • Wiki โ€“ https://github.com/ultralytics/yolov5/wiki
  • Tutorials โ€“ https://github.com/ultralytics/yolov5#tutorials
  • Docs โ€“ https://docs.ultralytics.com

Access additional Ultralytics โšก resources:

  • Ultralytics HUB โ€“ https://ultralytics.com
  • Vision API โ€“ https://ultralytics.com/yolov5
  • About Us โ€“ https://ultralytics.com/about
  • Join Our Team โ€“ https://ultralytics.com/work
  • Contact Us โ€“ https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 ๐Ÿš€ and Vision AI โญ!

cyberFoxi

comment created time in 6 hours

issue commentultralytics/yolov5

About Weight Initialization Technique

๐Ÿ‘‹ Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 ๐Ÿš€ resources:

  • Wiki โ€“ https://github.com/ultralytics/yolov5/wiki
  • Tutorials โ€“ https://github.com/ultralytics/yolov5#tutorials
  • Docs โ€“ https://docs.ultralytics.com

Access additional Ultralytics โšก resources:

  • Ultralytics HUB โ€“ https://ultralytics.com
  • Vision API โ€“ https://ultralytics.com/yolov5
  • About Us โ€“ https://ultralytics.com/about
  • Join Our Team โ€“ https://ultralytics.com/work
  • Contact Us โ€“ https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 ๐Ÿš€ and Vision AI โญ!

debparth

comment created time in 6 hours

pull request commentPyTorchLightning/pytorch-lightning

Hyperparameters for datamodule

Is this PR still in development?

tilman151

comment created time in 8 hours

issue commentultralytics/yolov5

More Augmentations

@glenn-jocher I am also interested in this question. As I understand it, YOLO slices images for training, but does it also know how to rotate images by (90, 180, 270) degrees? (And is it even worth doing that?)

jaideep11061982

comment created time in 11 hours

issue commentPyTorchLightning/pytorch-lightning

Memory explodes when limit_train_batches argument used

Hi @EdwardJB thanks for the issue! This does sound like unexpected behaviour. Could you try installing from master with pip install https://github.com/PyTorchLightning/pytorch-lightning/archive/master.zip and see if it persists? If so, then a notebook which reproduces the issue would be much appreciated.

Thanks :smiley:

EdwardJB

comment created time in 12 hours

issue commentPyTorchLightning/pytorch-lightning

consistency in logging images between multiple logging backends -- wandb and tensorboard

Hi @00krishna thanks for reporting this issue. It may also he related to #6720 and #7920 - please feel free to shar your thoughts on those as well.

Thanks :smiley:

00krishna

comment created time in 12 hours

issue commentultralytics/yolov5

Getting Segmentation Fault Error(core dumped) after successfully installing the coco datasets

@glenn-jocher Haven't heard anything from you. Please let me know

jaskiratsingh2000

comment created time in 12 hours

issue commentultralytics/yolov5

More Augmentations

@jaideep11061982 what augmentations are you interested in?

jaideep11061982

comment created time in 12 hours

issue commentultralytics/yolov5

How can I use DETRAC Dataset with the python3 test.py command?

@glenn-jocher Are you around and can answer how can this be possible?

jaskiratsingh2000

comment created time in 12 hours

push eventultralytics/yolov5

Glenn Jocher

commit sha b943a4b8507018f39c662caba9781b5fc105c89b

Update train.py

view details

push time in 12 hours

PR opened ultralytics/yolov5

Update train.py explicit arguments

@AyushExel this PR is for passing opt explicitly to train(). It looks like the logging may require significant rework, can you take a look at this please? Thanks!

+34 -33

0 comment

1 changed file

pr created time in 12 hours

create barnchultralytics/yolov5

branch : glenn-jocher-patch-1

created branch time in 12 hours

issue commentultralytics/yolov5

More Augmentations

๐Ÿ‘‹ Hello @jaideep11061982, thank you for your interest in ๐Ÿš€ YOLOv5! Please visit our โญ๏ธ Tutorials to get started, where you can find quickstart guides for simple tasks like Custom Data Training all the way to advanced concepts like Hyperparameter Evolution.

If this is a ๐Ÿ› Bug Report, please provide screenshots and minimum viable code to reproduce your issue, otherwise we can not help you.

If this is a custom training โ“ Question, please provide as much information as possible, including dataset images, training logs, screenshots, and a public link to online W&B logging if available.

For business inquiries or professional support requests please visit https://www.ultralytics.com or email Glenn Jocher at glenn.jocher@ultralytics.com.

Requirements

Python 3.8 or later with all requirements.txt dependencies installed, including torch>=1.7. To install run:

$ pip install -r requirements.txt

Environments

YOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

  • Google Colab and Kaggle notebooks with free GPU: <a href="https://colab.research.google.com/github/ultralytics/yolov5/blob/master/tutorial.ipynb">Open In Colab</a> <a href="https://www.kaggle.com/ultralytics/yolov5">Open In Kaggle</a>
  • Google Cloud Deep Learning VM. See GCP Quickstart Guide
  • Amazon Deep Learning AMI. See AWS Quickstart Guide
  • Docker Image. See Docker Quickstart Guide <a href="https://hub.docker.com/r/ultralytics/yolov5">Docker Pulls</a>

Status

CI CPU testing

If this badge is green, all YOLOv5 GitHub Actions Continuous Integration (CI) tests are currently passing. CI tests verify correct operation of YOLOv5 training (train.py), testing (test.py), inference (detect.py) and export (export.py) on MacOS, Windows, and Ubuntu every 24 hours and on every commit.

jaideep11061982

comment created time in 12 hours

issue openedultralytics/yolov5

More Augmentations

โ”Question

How i can use more and more augmentations related to Image Enhancements.

Additional context

created time in 12 hours