{"id":20391873,"url":"https://github.com/radiantearth/model_ramp_baseline","last_synced_at":"2025-04-12T11:42:06.258Z","repository":{"id":84825504,"uuid":"529028338","full_name":"radiantearth/model_ramp_baseline","owner":"radiantearth","description":"Replicable AI for Microplanning (Ramp) Bootstrap Model","archived":false,"fork":false,"pushed_at":"2023-09-25T09:34:00.000Z","size":19384,"stargazers_count":7,"open_issues_count":1,"forks_count":7,"subscribers_count":3,"default_branch":"main","last_synced_at":"2025-03-26T06:23:59.146Z","etag":null,"topics":["earth-observation","machine-learning","tensorflow"],"latest_commit_sha":null,"homepage":"https://mlhub.earth/model/model_ramp_baseline_v1","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/radiantearth.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null}},"created_at":"2022-08-25T21:49:03.000Z","updated_at":"2025-02-25T10:08:18.000Z","dependencies_parsed_at":"2023-09-29T02:25:37.275Z","dependency_job_id":null,"html_url":"https://github.com/radiantearth/model_ramp_baseline","commit_stats":{"total_commits":5,"total_committers":1,"mean_commits":5.0,"dds":0.0,"last_synced_commit":"89b799e764d76e5004fcb596a3ae4eda3dcc171d"},"previous_names":[],"tags_count":1,"template":false,"template_full_name":"radiantearth/mlhub_model_template","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/radiantearth%2Fmodel_ramp_baseline","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/radiantearth%2Fmodel_ramp_baseline/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/radiantearth%2Fmodel_ramp_baseline/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/radiantearth%2Fmodel_ramp_baseline/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/radiantearth","download_url":"https://codeload.github.com/radiantearth/model_ramp_baseline/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248563606,"owners_count":21125312,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["earth-observation","machine-learning","tensorflow"],"created_at":"2024-11-15T03:37:22.429Z","updated_at":"2025-04-12T11:42:06.230Z","avatar_url":"https://github.com/radiantearth.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Replicable AI for Microplanning (Ramp) Bootstrap Model\n\nThe Replicable AI for Microplanning (Ramp) deep learning model is a semantic\nsegmentation one which detects buildings from satellite imagery and delineates\nthe footprints in low-and-middle-income countries (LMICs) using satellite\nimagery and enables in-country users to build their own deep learning models\nfor their regions of interest. The architecture and approach were inspired by\nthe Eff-UNet model outlined in this\n[CVPR 2020 Paper](https://openaccess.thecvf.com/content_CVPRW_2020/papers/w22/Baheti_Eff-UNet_A_Novel_Architecture_for_Semantic_Segmentation_in_Unstructured_Environment_CVPRW_2020_paper.pdf).\n\n![model_ramp_baseline_v1](https://radiantmlhub.blob.core.windows.net/frontend-ml-model-images/model_ramp_baseline_v1.png)\n\nMLHub model id: `model_ramp_baseline_v1`. Browse on [Radiant MLHub](https://mlhub.earth/model/model_ramp_baseline_v1).\n\n## ML Model Documentation\n\nPlease review the model architecture, license, applicable spatial and temporal extents\nand other details in the [model documentation](/docs/index.md).\n\n## System Requirements\n\n* Git client\n* [Docker](https://www.docker.com/) with\n    [Compose](https://docs.docker.com/compose/) v1.28 or newer.\n\n## Hardware Requirements\n\n|            |Inferencing|Training|\n|------------|-----------|--------|\n| RAM        | 4 GB RAM    | [View Ramp model card](https://rampml.global/ramp-model-card/) |\n| NVIDIA GPU | optional  | [required](https://rampml.global/ramp-model-card/) |\n\n## Get Started With Inferencing\n\nFirst clone this Git repository. Please note: this repository uses\n[Git Large File Support (LFS)](https://git-lfs.github.com/) to include the\nmodel checkpoint file. Either install `git lfs` support for your git client,\nuse the official Mac or Windows GitHub client to clone this repository.\n\n```bash\ngit clone https://github.com/radiantearth/model_ramp_baseline.git\ncd model_ramp_baseline/\n```\n\nAfter cloning the model repository, you can use the Docker Compose runtime\nfiles as described below.\n\nPlease note: these command-line examples were tested on Linux and MacOS.\nWindows and WSL users may need to substitute appropriate commands, especially\nfor setting environment variables.\n\n## Pull or Build the Docker Image\n\nPull pre-built image from Docker Hub (recommended):\n\n```bash\n# cpu\ndocker pull docker.io/radiantearth/model_ramp_baseline:1\n# optional, for NVIDIA gpu\ndocker pull docker.io/radiantearth/model_ramp_baseline:1-gpu\n\n```\n\nOr build image from source:\n\n```bash\n# cpu\ndocker build -t radiantearth/model_ramp_baseline:1 -f Dockerfile_cpu .\n# for NVIDIA gpu\ndocker build -t radiantearth/model_ramp_baseline:1-gpu -f Dockerfile_gpu .\n\n```\n\n## Run Model to Generate New Inferences\n\n1. Prepare your input and output data folders. The `data/` folder in this repository\n    contains some placeholder files to guide you.\n\n    * The `data/` folder must contain:\n        * `input/chips` imagery chips for inferencing.\n            [For example, Maxar ODP imagery](https://rampml.global/ramp-faqs/)\n            * File name: `chip_id.tif` for example:\n                `0fec2d30-882a-4d1d-a7af-89dac0198327.tif`.\n            * File Format: GeoTIFF, 256x256\n            * Coordinate Reference System: WGS84, EPSG:4326\n            * Bands: 3 bands per file:\n                * Band 1 Type=Byte, ColorInterp=Red\n                * Band 2 Type=Byte, ColorInterp=Green\n                * Band 3 Type=Byte, ColorInterp=Blue\n        * `/input/checkpoint.tf` the model checkpoint folder in tensorflow format.\n            Please note: the model checkpoint is included in this repository.\n    * The `output/` folder is where the model will write inferencing results.\n\n2. Set `INPUT_DATA` and `OUTPUT_DATA` environment variables corresponding with\n    your input and output folders. These commands will vary depending on operating\n    system and command-line shell:\n\n    ```bash\n    # change paths to your actual input and output folders\n    export INPUT_DATA=\"/home/my_user/model_ramp_baseline/data/input/\"\n    export OUTPUT_DATA=\"/home/my_user/model_ramp_baseline/data/output/\"\n    ```\n\n3. Run the appropriate Docker Compose command for your system:\n\n    Use either `docker compose` or `docker-compose` depending on your system.\n\n    ```bash\n    # cpu\n    docker compose up model_ramp_baseline_v1_cpu\n    # NVIDIA gpu driver\n    docker compose up model_ramp_baseline_v1_gpu\n    ```\n\n4. Wait for the `docker compose` to finish running, then inspect the\n`OUTPUT_DATA` folder for results.\n\n## Understanding Output Data\n\nPlease review the model output format and other technical details in the [model\ndocumentation](/docs/index.md).\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fradiantearth%2Fmodel_ramp_baseline","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fradiantearth%2Fmodel_ramp_baseline","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fradiantearth%2Fmodel_ramp_baseline/lists"}