Visual R-CNN model training studio

R-CNN TRAINER

Build. Train. Detect. Export.

Dark desktop workflow for custom R-CNN detectors, visual backbone design, large geospatial datasets and expert CUDA/CPU training.

Visual backbone builderFaster R-CNN / Mask R-CNNLarge image chunkingONNX export

Open Microsoft Store Direct Microsoft install

Train R-CNN models with full control

R-CNN Trainer is made for users who need configurable detector training instead of a black-box pipeline: datasets, bands, backbone, anchors, epochs, CUDA/CPU and export steps are all exposed.

Faster R-CNN and Mask R-CNN style training workflow
Embedded visual backbone studio with draggable model blocks
Dataset/project preparation with image chunking for huge maps
Training bands for RGB, false color or custom satellite inputs
Empty/negative chunks support for better false-positive control
CUDA or CPU training selection with mixed precision options
Checkpoint resume, final checkpoint export and ONNX export
Built-in Python console for custom training and inference commands

Workflow

From dataset to trained detector in a controlled desktop pipeline.

Create project

Prepare folders, images, labels, annotations, exports, weights and project metadata.

Prepare dataset

Chunk very large imagery, keep empty negatives and define the bands used for training.

Build backbone

Use the visual graph creator to design a custom CNN/FPN backbone for R-CNN models.

Train and export

Train with selected parameters, resume checkpoints and export final models to ONNX.

Screenshots

Orange/black Store-style visuals focused on the R-CNN Trainer interface and large-image detection workflow.

R-CNN Trainer visual backbone studio — Visual R-CNN backbone studio with block graph, training parameters and output settings.

R-CNN large image detection interface — Large-image tiled R-CNN detection with filter slider, NMS, class display and GIS export controls.

Downloads

The main R-CNN Trainer links are preserved and adapted from the existing YOLO Trainer HTML structure.

Microsoft Store

Open the R-CNN Trainer product page in the Microsoft Store.

Direct Microsoft installer

Start the Microsoft App Installer flow directly.

Install

YOLO Trainer

Related trainer app for YOLO workflows.

Open YOLO Trainer

RF-DETR Trainer

Related trainer app for RF-DETR workflows.

Open RF-DETR Trainer

Designed in the same structure as the YOLO Trainer page, but rewritten for the orange/black R-CNN Trainer branding.

Other trainers

More AI and geospatial tools from the same desktop workflow family.

YOLO Trainer

Train, annotate and run YOLO detection workflows.

RF-DETR Trainer

Transformer-based object detection trainer for advanced experiments.

Mustatil

Large-image geospatial AI suite for detection, annotation and GIS export.