How xiaobul/cm-unet works
CM-UNet is a remote-sensing semantic-segmentation research application for LoveDA, Vaihingen, Potsdam and UAVid imagery. The main workflow configures a dataset and model, trains with segmentation losses and validation metrics, then predicts masks for tiled or large imagery. The repository also contains image-preparation utilities and several segmentation architectures. The graph combines sampled call sites and SOURCE INDEX references; links for unsampled capabilities are omitted where their wiring cannot be established.
Main entry points: train_supervision.py Training runner · vision_mamba.py Mamba variants · tools Dataset preparation · mambaunet.py Dataset profiles
Components 22
Data preparation
- LoveDA dataset
cmunet/geoseg/datasets/loveda_dataset.py - Vaihingen dataset
cmunet/geoseg/datasets/vaihingen_dataset.py - Potsdam dataset
cmunet/geoseg/datasets/potsdam_dataset.py - UAVid dataset
cmunet/geoseg/datasets/uavid_dataset.py - Image transforms
cmunet/geoseg/datasets/transform.py - Dataset preparation
cmunet/tools
Segmentation models
- UNetFormer
cmunet/geoseg/models/UNetFormer.py - Mamba variants
cmunet/geoseg/models/networks/vision_mamba.py - Mamba backbone
cmunet/geoseg/models/networks/mamba_sys.py - Residual Mamba
cmunet/geoseg/models/networks/mamba_sys_res.py - Swin Mamba
cmunet/geoseg/models/networks/mamba_sys_swin.py - VMamba backbone
cmunet/geoseg/models/networks/mamba_sys_vmamba.py - Selective scan
cmunet/geoseg/Mamba-UNet/mamba/mamba_ssm/ops/selective_scan_interface.py
Training and evaluation
- Run configuration
cmunet/config.py - Dataset profiles
cmunet/config/loveda/mambaunet.py - Training runner
cmunet/train_supervision.py - Segmentation losses
cmunet/geoseg/losses - Segmentation metrics
cmunet/tools/metric.py
Inference
- Large-image inference
cmunet/inference_huge_image.py - UAVid inference
cmunet/inference_uavid.py - Inference tiles
cmunet/inference_uavid.py
Other components
- Researcher
Connections 28
- Researcher → Training runner: starts training
- LoveDA dataset → Image transforms: uses transforms
cmunet/geoseg/datasets/loveda_dataset.py - Vaihingen dataset → Image transforms: uses transforms
cmunet/geoseg/datasets/vaihingen_dataset.py - Potsdam dataset → Image transforms: uses transforms
cmunet/geoseg/datasets/potsdam_dataset.py - UAVid dataset → Image transforms: uses transforms
cmunet/geoseg/datasets/uavid_dataset.py - Dataset profiles → LoveDA dataset: configures data
cmunet/config/loveda/mambaunet.py - Dataset profiles → UNetFormer: imports model
cmunet/config/loveda/mambaunet.py - Dataset profiles → Mamba variants: constructs model
cmunet/config/loveda/mambaunet.py - Dataset profiles → Segmentation losses: selects loss
cmunet/config/loveda/mambaunet.py - Training runner → Run configuration: loads settings
cmunet/train_supervision.py - Training runner → Segmentation metrics: updates metrics
cmunet/train_supervision.py - Training runner → Mamba variants: runs network
- Training runner → Segmentation losses: computes loss
- Mamba variants → Mamba backbone: selects backbone
cmunet/geoseg/models/networks/vision_mamba.py - Mamba variants → Residual Mamba: selects backbone
cmunet/geoseg/models/networks/vision_mamba.py - Mamba variants → Swin Mamba: selects backbone
cmunet/geoseg/models/networks/vision_mamba.py - Mamba variants → VMamba backbone: selects backbone
cmunet/geoseg/models/networks/vision_mamba.py - VMamba backbone → Selective scan: calls scan
cmunet/geoseg/models/networks/mamba_sys_vmamba.py - Large-image inference → Run configuration: loads config
cmunet/inference_huge_image.py - Large-image inference → Training runner: loads checkpoint
cmunet/inference_huge_image.py - UAVid inference → Run configuration: loads config
cmunet/inference_uavid.py - UAVid inference → Training runner: loads checkpoint
cmunet/inference_uavid.py - UAVid inference → Segmentation metrics: evaluates masks
cmunet/inference_uavid.py - UAVid inference → Inference tiles: batches tiles
cmunet/inference_uavid.py - Dataset preparation → LoveDA dataset: prepares masks
- Dataset preparation → Vaihingen dataset: splits patches
- Dataset preparation → Potsdam dataset: splits patches
- Dataset preparation → UAVid dataset: splits patches