fine tuning on custom dataset

IDEA-Research / DINO

[ICLR 2023] Official implementation of the paper "DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection"

Apache License 2.0

2.08k stars 228 forks source link

fine tuning on custom dataset #234

Open zyx20180209 opened 8 months ago

zyx20180209 commented 8 months ago

I tried to fine tune one pre trained model using another dataset, after that I want to use more training data to boost the performance. When I rerun bash scripts/DINO_train.sh it finished the process instantly and it seems the model keep unchanged. How could I fine tune the same pre trained model with another dataset?

half-truism commented 8 months ago

Read the readme, section 6 under run https://github.com/IDEA-Research/DINO/tree/main#run

zyx20180209 commented 7 months ago

I followed what the instruction said, the fine tuning worked when I did it for the first time. The problem is after that the model seems fixed and cannot be further train or fine-tuned.