ucbdrive/few-shot-object-detection
Few-shot object detection framework implementing the TFA two-stage fine-tuning approach from ICML 2020.
Not currently ranked — collecting fresh signals.
star history
FsDet provides implementations for few-shot object detection benchmarks on PASCAL VOC, COCO, and LVIS datasets. The repository includes a two-stage fine-tuning approach where an object detector is first trained on base classes with abundant data, then only the last layers are fine-tuned on a small balanced dataset of novel classes. It offers benchmark results, pre-trained models, and modularized code for adding custom datasets and models.
Frequently asked
- What is ucbdrive/few-shot-object-detection?
- Few-shot object detection framework implementing the TFA two-stage fine-tuning approach from ICML 2020.
- Is few-shot-object-detection open source?
- Yes — ucbdrive/few-shot-object-detection is open source, released under the Apache-2.0 license.
- What language is few-shot-object-detection written in?
- ucbdrive/few-shot-object-detection is primarily written in Python.
- How popular is few-shot-object-detection?
- ucbdrive/few-shot-object-detection has 1.2k stars on GitHub.
- Where can I find few-shot-object-detection?
- ucbdrive/few-shot-object-detection is on GitHub at https://github.com/ucbdrive/few-shot-object-detection.