Home - byuawsfhtl/RLL_computer_vision GitHub Wiki
skip to supercomputer paths
skip to Vdrive paths
Welcome
Welcome to the computer-vision wiki!
This wiki shares knowledge for everything you will need to succeed on our team and beyond. You can navigate the sidebar to find information on whatever topic you need. If you don't see a page for a topic you think would be helpful, feel free to add it!
If you are new to the team or want to build up fundamental programming or ML skills, there are tutorials, cheat sheets, and readings on the Getting Started page.
Below are paths to the data for our current projects.
Using GitHub and the RLL Organization
If you are not familiar with using git or GitHub, review this quickstart or this post for CLI. If videos are more your thing, watch this YouTube tutorial. Many IDEs have native version control compatible with GitHub.
When you start a new project or continue a current project, please use repositories in the computer-vision team (not the RLL organization or the repository housing this wiki) and push your changes to a branch there if at all possible. This allows other RAs to access the code, see your changes, and make their own contributions in a centralized location that saves space in the Vdrive and the supercomputer. It is also much clearer what work is done and what work needs to be finished when there are turnovers of RAs.
If you don't have access to the team for some reason, please contact [email protected] to be added.
When creating a repository, please add a README.md that describes the project, its goals, and locations for data. It should effectively present the project.
Suggested GitHub tools
- Issues: if there is a specific problem you are working on or find a bug in the code, you can add an issue to let anyone interested in that project know of the problem. There is space for conversations about it and someone who might be more experienced with the problem could get to it quicker than if we talk about it in a team meeting.
- Codespaces are integrated IDEs in GitHub that allow you to develop code with less interaction with an environment. You can run Jupyter, Django, and others straight from GitHub. This is good if you don't have resources on your computer for your project.
- Projects: your issues can be added to a project (or project steps added as issues) to help streamline assignments for each project. In the image below we get an idea of what has been done on the project and what is left to do so that other RAs have a sense of what you are doing. It is also helpful when you are setting goals for each stage of a project.
~/ Indicates a path from your home directory (/fslhome/username)
Supercomputer paths:
Path | Description |
---|---|
~/fsl_groups/fslg_JoePriceResearch/compute/tesseract_ocr_env | environment path for Tesseract wrappers (Tesserocr and Pytesseract) |
~/fsl_groups/fslg_census/compute/projects/Mexico_Census | Home path of Mexico Census project |
~/fsl_groups/fslg_census/compute/projects/Mexico_Census/language2 | Language2 classification |
~/fsl_groups/fslg_census/compute/projects/Mexico_Census/language2/class2Env | Environment that can run the HWR training/inference |
~/fsl_groups/fslg_death/compute/utah_death_2 | Home path of the Utah Death Records Project |
~/fsl_groups/fslg_census/compute/projects/colorado_patents | Home path of the Colorado Land Patents Project |
~/fsl_groups/fslg_census/compute/projects/1920_headers | Home path of the 1920 headers project |
~/fsl_groups/fslg_census/Mexico_Census/HWR | handwriting model |
~/fsl_groups/fslg_census/hwr/trained_models | handwriting model |
~/fsl_groups/fslg_handwriting/compute/generalized_hwr | handwriting model |
~/fsl_groups/fslg_handwriting/simple_hwr | handwriting model |
~/fsl_groups/fslg_handwriting/old_simple_hwr | handwriting model |
~/fsl_groups/fslg_death/compute/hwr | handwriting model |
~/fsl_groups/fslg_census/compute/data_raw/fs_us_census_1940/records | 1940 Census raw data |
~/fsl_groups/fslg_census1940/compute/imgs/segmented/snippets | 1940 Census snippets |
~/fsl_groups/fslg_census1940/compute/denmark_hwr/predict-packages | Torben's predict packages (1940 Census); models for age, birthplace, country, first name, gender, incorporated place, last name, race, relationship, state |
~/fsl_groups/fslg_handwriting/compute/ww_project/hwr/hwr | Wilford Woodruf handwriting model |
V Drive paths:
Path | Description |
---|---|
V:\lab_items\HWR_predictions.mp4 | A video demonstrating how to use Demark's HWR models that they build |
Cool links:
AI models and datasets for training (hugging face)
Wiki page updates:
Tesseract on the Supercomputer has been moved to OCR How to mount the Supercomputer onto VS code has been moved to Supercomputer Basics Making Training Data Specifics has been moved to Creating Training Data Mapping the V drive to a Lab Computer has been moved to V Drive Sending Data for HWR has been moved to Denmark: Handwriting Training