Caleb Robinson calebrob6
calebrob6
/ ftw-test.ipynb
Created
November 18, 2025 06:23
Example running Fields of The World model with TorchGeo
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
calebrob6
/ dinov3_and_anyup.ipynb
Created
November 14, 2025 21:22
Notebook showing how to use the AnyUp model to upsample low (spatial) resolution embeddings.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| import argparse | |
| import math | |
| import os | |
| import time | |
| from typing import List, Optional, Sequence, Tuple | |
| import numpy as np | |
| import rasterio | |
| import rasterio.windows | |
| import torch |
calebrob6
/ FTW demo application.ipynb
Created
August 12, 2025 23:29
Notebook that shows end-to-end example of running a Fields of the World model on Sentinel-2 imagery and joining with CDL labels and MOSAIKS embeddings.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| def batch_histogram(data_tensor, num_classes=-1): | |
| """ | |
| From https://discuss.pytorch.org/t/batched-torch-histc/179741 | |
| Computes histograms of integral values, even if in batches (as opposed to torch.histc and torch.histogram). | |
| Arguments: | |
| data_tensor: a D1 x ... x D_n torch.LongTensor | |
| num_classes (optional): the number of classes present in data. | |
| If not provided, tensor.max() + 1 is used (an error is thrown if tensor is empty). | |
| Returns: |
calebrob6
/ benchmark_sieve.py
Created
April 17, 2025 19:15
Script for benchmarking polygonization through gdal command line calls vs. in-memory with rasterio and friends.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| #!/usr/bin/env python3 | |
| """ | |
| benchmark_sieve.py | |
| Benchmark sieving, polygonizing, and simplifying GeoTIFFs using two methods: | |
| - "gdal": subprocess calls to gdal_sieve.py, gdal_polygonize.py, and ogr2ogr | |
| - "python": pure Python using rasterio.features, shapely, and fiona | |
| """ | |
| import os |
calebrob6
/ rcf_segmentation.py
Created
March 13, 2025 03:26
A subclass of torchgeo's RCF model that averages features over a given mask instead of the whole input.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| class RCFSegmentationFeatures(RCF): | |
| def forward(self, x: torch.Tensor, y: torch.Tensor) -> torch.Tensor: | |
| """Forward pass of the RCF model. | |
| Args: | |
| x: a tensor with shape (C, H, W) | |
| y: a tensor with shape (H, W) | |
| Returns: | |
| a tensor of size (``self.num_features``) |
calebrob6
/ itu_global_unconnected.ipynb
Created
February 10, 2025 16:50
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
calebrob6
/ eurosat_few_shot.ipynb
Last active
January 13, 2025 22:09
Small experiment showing KNN performance on EuroSAT with MOSAIKS features and different numbers of training samples per class.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
NewerOlder