search

Asad-Ismail / Grounding-Dino-FineTuning

Fine tuning grounding Dino
Apache License 2.0
53 stars 8 forks source link

test the trained model #4

Open mengyu212 opened 5 months ago

mengyu212 commented 5 months ago

hello!thanks for your code, that's useful! However, I have several problems during the training process. 1) How much data did you use in fine-tuning？ 2) When I used the fine-tuned model to test, I got more than 800 detection boxes, but there were only 3 objects in the image actually. Have you met this situation? I hope you can help with this. Thank you again and best wishes!

Asad-Ismail commented 5 months ago

Hi @mengyu212 thank you for your question. We used about 2000 images to finetune the model. We had multiple detections for the same object but NMS implementation helped there it is also included in this repo.

vinaluna commented 5 months ago

Hi @mengyu212 thank you for your question. We used about 2000 images to finetune the model. We had multiple detections for the same object but NMS implementation helped there it is also included in this repo.

Thanks for your contribution! There are some problems I want to ask. 1)May I ask the time cost and hardware requirement (e.g. GPU) of this finetuning? 2)How many images should be used at least to finetune the model so that we can achieve the comparable performance as displayed in your test file? Hope you can help with these issues. Have a nice day!