Merged
Conversation
…ort multimodal dependencies.
Collaborator
Author
|
It looks like source code files weren't committed properly. There are multiple references to a @seohyun408, @seungjindes, @y00628, if any of you still have the source code, I'd appreciate if you push it. If not, I can try to reverse engineer based on the existing code. |
Collaborator
|
The original source code will likely be difficult to obtain. Even if it is
not a perfect reproduction, I think it would be best to reimplement it in a
simplified form based on the existing code. Thank you.원본 소스 코드를 얻기 어려울 수
있습니다. 완벽한 재현은 아니더라도 기존 코드를 기반으로 단순화된 형태로 다시 구현하는 것이 가장 좋을 것 같습니다. 감사합니다.
…________________________________________
______________________________________________________
Seungjin Han, M.S한승진, M.S.
Data eXperience Laboratory데이터 경험 연구실
Department of Applied Artificial Intelligence응용인공지능학과
Sungkyunkwan University성균관대학교
Seoul, Republic of Korea대한민국 서울
Tel: +82 010-6638-2302 /전화: +82 010-6638-2302 /
Email: ***@***.***이메일: ***@***.***
Web: https://seungjindes.github.io/about/ / Lab: http://dsl.skku.edu웹:
https://seungjindes.github.io/about/ / 연구실: http://dsl.skku.edu
________________________________________
______________________________________________________
2026년 2월 27일 (금) AM 10:37, Andrew Scouten ***@***.***>님이 작성:
*andrewscouten* left a comment (collaborativebioinformatics/OncoLearn#15)
<#15 (comment)>
It looks like source code files weren't committed properly. There are
multiple references to a src/multimodal/src/data module, but the
src/multimodal/.gitignore references /src/data/*. It's likely these files
were unintentionally ignored, but as a result the code cannot run as-is.
@seohyun408 <https://github.com/seohyun408>, @seungjindes
<https://github.com/seungjindes>, @y00628 <https://github.com/y00628>, if
any of you still have the source code, I'd appreciate if you push it. If
not, I can try to reverse engineer based on the existing code.
—
Reply to this email directly, view it on GitHub
<#15 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/BE3NC6INKTQZ3ICYZT4VAJ34OCFG5AVCNFSM6AAAAACWA5VZX6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTSNZUGQ2TIMRSG4>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
Collaborator
|
I have uploaded the preprocessing code used for Andrew’s training as well
as the README file. Although I could not find the original raw data files,
I hope these materials will still be helpful.
Also, regarding the hackathon paper, would it be okay if I make some edits
to the draft we have written?
Thank you.
…________________________________________
Seungjin Han, M.S
Data eXperience Laboratory
Department of Applied Artificial Intelligence
Sungkyunkwan University
Seoul, Republic of Korea
Tel: +82 010-6638-2302 /
Email: ***@***.***
Web: https://seungjindes.github.io/about/ / Lab: http://dsl.skku.edu
________________________________________
2026년 2월 27일 (금) PM 3:14, Seungjin Han ***@***.***>님이 작성:
The original source code will likely be difficult to obtain. Even if it is
not a perfect reproduction, I think it would be best to reimplement it in a
simplified form based on the existing code. Thank you.원본 소스 코드를 얻기 어려울 수
있습니다. 완벽한 재현은 아니더라도 기존 코드를 기반으로 단순화된 형태로 다시 구현하는 것이 가장 좋을 것 같습니다. 감사합니다.
________________________________________
______________________________________________________
Seungjin Han, M.S한승진, M.S.
Data eXperience Laboratory데이터 경험 연구실
Department of Applied Artificial Intelligence응용인공지능학과
Sungkyunkwan University성균관대학교
Seoul, Republic of Korea대한민국 서울
Tel: +82 010-6638-2302 /전화: +82 010-6638-2302 /
Email: ***@***.***이메일: ***@***.***
Web: https://seungjindes.github.io/about/ / Lab: http://dsl.skku.edu웹:
https://seungjindes.github.io/about/ / 연구실: http://dsl.skku.edu
________________________________________
______________________________________________________
2026년 2월 27일 (금) AM 10:37, Andrew Scouten ***@***.***>님이 작성:
> *andrewscouten* left a comment (collaborativebioinformatics/OncoLearn#15)
> <#15 (comment)>
>
> It looks like source code files weren't committed properly. There are
> multiple references to a src/multimodal/src/data module, but the
> src/multimodal/.gitignore references /src/data/*. It's likely these
> files were unintentionally ignored, but as a result the code cannot run
> as-is.
>
> @seohyun408 <https://github.com/seohyun408>, @seungjindes
> <https://github.com/seungjindes>, @y00628 <https://github.com/y00628>,
> if any of you still have the source code, I'd appreciate if you push it. If
> not, I can try to reverse engineer based on the existing code.
>
> —
> Reply to this email directly, view it on GitHub
> <#15 (comment)>,
> or unsubscribe
> <https://github.com/notifications/unsubscribe-auth/BE3NC6INKTQZ3ICYZT4VAJ34OCFG5AVCNFSM6AAAAACWA5VZX6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTSNZUGQ2TIMRSG4>
> .
> You are receiving this because you were mentioned.Message ID:
> ***@***.***>
>
|
Collaborator
|
The content to be included in the paper has been uploaded to Discord by
River.
…________________________________________
Seungjin Han, M.S
Data eXperience Laboratory
Department of Applied Artificial Intelligence
Sungkyunkwan University
Seoul, Republic of Korea
Tel: +82 010-6638-2302 /
Email: ***@***.***
Web: https://seungjindes.github.io/about/ / Lab: http://dsl.skku.edu
________________________________________
2026년 2월 27일 (금) PM 6:50, Seungjin Han ***@***.***>님이 작성:
I have uploaded the preprocessing code used for Andrew’s training as well
as the README file. Although I could not find the original raw data files,
I hope these materials will still be helpful.
Also, regarding the hackathon paper, would it be okay if I make some edits
to the draft we have written?
Thank you.
________________________________________
Seungjin Han, M.S
Data eXperience Laboratory
Department of Applied Artificial Intelligence
Sungkyunkwan University
Seoul, Republic of Korea
Tel: +82 010-6638-2302 /
Email: ***@***.***
Web: https://seungjindes.github.io/about/ / Lab: http://dsl.skku.edu
________________________________________
2026년 2월 27일 (금) PM 3:14, Seungjin Han ***@***.***>님이 작성:
> The original source code will likely be difficult to obtain. Even if it
> is not a perfect reproduction, I think it would be best to reimplement it
> in a simplified form based on the existing code. Thank you.원본 소스 코드를 얻기 어려울
> 수 있습니다. 완벽한 재현은 아니더라도 기존 코드를 기반으로 단순화된 형태로 다시 구현하는 것이 가장 좋을 것 같습니다. 감사합니다.
> ________________________________________
> ______________________________________________________
> Seungjin Han, M.S한승진, M.S.
>
> Data eXperience Laboratory데이터 경험 연구실
> Department of Applied Artificial Intelligence응용인공지능학과
> Sungkyunkwan University성균관대학교
> Seoul, Republic of Korea대한민국 서울
>
> Tel: +82 010-6638-2302 /전화: +82 010-6638-2302 /
> Email: ***@***.***이메일: ***@***.***
> Web: https://seungjindes.github.io/about/ / Lab: http://dsl.skku.edu웹:
> https://seungjindes.github.io/about/ / 연구실: http://dsl.skku.edu
> ________________________________________
> ______________________________________________________
>
>
> 2026년 2월 27일 (금) AM 10:37, Andrew Scouten ***@***.***>님이
> 작성:
>
>> *andrewscouten* left a comment
>> (collaborativebioinformatics/OncoLearn#15)
>> <#15 (comment)>
>>
>> It looks like source code files weren't committed properly. There are
>> multiple references to a src/multimodal/src/data module, but the
>> src/multimodal/.gitignore references /src/data/*. It's likely these
>> files were unintentionally ignored, but as a result the code cannot run
>> as-is.
>>
>> @seohyun408 <https://github.com/seohyun408>, @seungjindes
>> <https://github.com/seungjindes>, @y00628 <https://github.com/y00628>,
>> if any of you still have the source code, I'd appreciate if you push it. If
>> not, I can try to reverse engineer based on the existing code.
>>
>> —
>> Reply to this email directly, view it on GitHub
>> <#15 (comment)>,
>> or unsubscribe
>> <https://github.com/notifications/unsubscribe-auth/BE3NC6INKTQZ3ICYZT4VAJ34OCFG5AVCNFSM6AAAAACWA5VZX6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTSNZUGQ2TIMRSG4>
>> .
>> You are receiving this because you were mentioned.Message ID:
>> ***@***.***>
>>
>
|
Collaborator
Author
|
@seungjindes I am not the team's writer. I would ask one of the two of them, if they have not changed since. I am of the opinion that this is okay, but I would make sure first. |
…dules Add Multimodal's README.md so that refactor can reference it more easily.
…learn-modules [Refactor] Oncolearn
…rs while keeping original functionality. Still needs more testing.
- Added a new `encoders.py` file to implement an encoder registration system. - Implemented `register_encoder`, `get_encoder`, and `get_all_encoders` functions for managing encoders. - Updated imports in `train_nvflare.py` to reflect new encoder structure. - Refactored `trainer.py` to utilize PyTorch Lightning for training orchestration. - Removed legacy configuration management code from `config.py`. - Added tests for configuration loading and validation in `test_config.py`. - Created example configuration files for testing purposes.
- Added YAML error handling in XenaCohortBuilder to raise ValueError for invalid configurations. - Filtered empty cohort names in download script to prevent processing errors. - Initialized _full_dataset in ImageDataModule and ClinicalDataModule to improve data handling. - Updated PillowLoader to provide more informative error messages for image loading failures. - Improved dataset validation in MultimodalDataModule to ensure only valid labels are processed. - Enhanced encoder classes to conditionally freeze models based on configuration settings.
- Enhanced OncoTrainer to support hyperparameter optimization (HPO) using Optuna, including a new method to run HPO and apply best parameters to the training configuration. - Updated L1 regularization calculation in BaseOncoClassifier to only include parameters that require gradients. - Changed the registration string for GatedLateFusionConfig to include the full module path. - Adjusted gradient clipping value handling in OncoTrainer to ensure it is only applied when greater than zero. - Updated dependency management in `uv.lock` to include new packages: alembic, colorlog, greenlet, mako, optuna, and sqlalchemy, along with their respective versions and dependencies.
…ic assay data; enhance API client with retry logic
…zer and loss parameter handling in YAML and code
…function flexibility
…d later to "data/source"
- Introduced unit tests for the pipeline executor in `test_pipeline_executor.py`, covering various scenarios including loading data, joining datasets, and handling errors. - Added unit tests for pipeline nodes in `test_pipeline_nodes.py`, validating default behaviors and configurations for `DataSource`, `Load`, `Join`, `Sequence`, and modality classes. - Refactored image and multimodal data modules to improve structure and consistency in `test_image_e2e.py`, `test_multimodal_e2e.py`, and `test_tabular_e2e.py`. - Updated configuration tests in `test_config.py` to reflect changes in the pipeline-based schema and removed deprecated modality tests. - Consolidated data module tests in `test_datamodules.py` to focus on the new `ImageDataModule` and removed legacy tests for `GeneDataModule` and `ClinicalDataModule`. - Enhanced the dataset registry tests in `test_registry.py` to include dataset registration and retrieval functionalities.
…e labels; refactor data modules and add Log2Normalization support
- Created train and test split files for fold 0 to fold 4 in the PAM50 and stage datasets. - Implemented logging functionality to capture the KFold generation process, including patient counts and splits. - Updated Docker Compose configuration to mount the configs directory for easier access within containers. - Enhanced the kfold.py script to log output to a file while also displaying it in the console.
…metrics logging callback
…ctions and key capabilities section
Collaborator
Author
|
While final verification still needs to be run to ensure it meets the described metrics and methods in the paper... I will have to take an extended break for exams. As our manuscript will be published soon, I want to push this out before hand. The current code in this PR:
As such, I am going to be merging this branch to main so that the project is in a more complete state when the manuscript releases. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Refactors multimodal code to work with:
Closes #5, #13