@naxatra
Nothing here yet.
Nothing here yet.
Aug 31, 2025 · 2 min read · After spending the previous weeks building and testing an active learning pipeline (mostly inside Kaggle notebooks), I dedicated Weeks 9–10 to refinement and usability. This meant transforming rough experiments into something maintainable, reusable, ...
Join discussion
Aug 31, 2025 · 4 min read · In my last week’s blog, I had a working active learning pipeline: a loop where RetinaNet50 was trained on a labeled pool, predictions were run on an unlabeled pool, uncertain samples were selected, and the model was retrained with those added samples...
Join discussion
Aug 31, 2025 · 3 min read · In the previous two weeks, I focused entirely on fine-tuning RetinaNet50 on custom datasets and establishing a baseline. That work gave me a good sense of how well the model adapts to new data and what pitfalls to expect during training. With that fo...
Join discussion
Aug 31, 2025 · 2 min read · In Weeks 3–4, my project shifted direction. While the initial plan emphasized implementing multiple active learning sampling strategies and creating a CLI script, mentor discussions guided me toward a more impactful path — fine-tuning RetinaNet50 on ...
Join discussion
Jun 20, 2025 · 3 min read · In these two weeks my work focused on adding an Active Learning module to DeepForest to enable efficient selection of image tiles for annotation. The aim is to provide a framework that selects the most informative samples, deferring integration with ...
Join discussion