Package index • petrographer

All functions

evaluate_model_sahi(): Evaluate detections with SAHI and COCO metrics

evaluate_training()

Evaluate model training Reads Detectron2 metrics.json and exports:

training_metrics.csv: losses, lr, etc.
validation_metrics.csv: aggregate COCO bbox and segm AP metrics
validation_classwise.csv: per-class AP metrics (when logged by evaluator)

from_pretrained(): Load a pretrained model

get_dataset_path(): Get path to pinned dataset

get_population_stats(): Get overall population statistics

get_training_dataset(): Get the dataset used to train a model

list_datasets(): List pinned datasets

list_models(): List available models

list_trained_models(): List all locally trained models

model_info(): Get model info

pg_plot_annotations(): Plot COCO annotations on an image

pg_plot_dataset_image(): Plot a sample image from a dataset directory

pg_save_dataset_preview(): Save a preview image for a dataset

pin_dataset(): Pin a dataset to a board

pin_model(): Pin a trained model to a board

predict(): Predict objects in an image

predict_image(): Predict objects in a single image

predict_images(): Predict objects in multiple images (directory)

slice_dataset(): Slice COCO dataset for varying image sizes

summarize_by_image(): Summarize detections by image

train_model(): Train a new petrography detection model

validate_dataset(): Validate a COCO-style dataset directory