devit
devit copied to clipboard
mlzxy
1. I encountered torch.cuda.OutOfMemoryError: CUDA out of memory error, but I used 8 3090 graphics cards and did not encounter it when evaluating the lvis dataset, I tried to reduce...
Hello， I am trying to run the vit/l . but the predicted results is lower than the results in paper. box APr in paper is 33.6 ,but the box APr...
I ran the demo.py script with the provided ycb_prototypes.pth prototype model file. However, I only used the first 10 prototypes for both label_names and test_class_weight (apple, ball, banana, bowl, brick,...
Excellent job, I want to use it for object detection in specific datasets, but how to prepare my own dataset or what steps should I follow? Thank you again！
I'm trying to run the demo on a different dataset and the inference speed is extremely slow. It takes ~2min/image to run the model on CPU and 5s/image on GPU....
Hi, really impressive work! I was just wondering if the config files in the repo contain the hyperparameters for the best model? For example, do the parameters in [configs/open-vocabulary/coco/vits.yaml](https://github.com/mlzxy/devit/blob/main/configs/open-vocabulary/coco/vits.yaml) give...
May I ask if there is any inference code provided and I would like to know the inference code for the feed shot.
This is a very nice and inspiring work! However, I found a bug which lead wrong mAP in few-shot evaluation. It is in file "devit/lib/data/fewshot.py", line 460. It will lead...
I need to test metric on my dataset, i follow the instruction ( https://github.com/mlzxy/devit/issues/4) to build CLASS_PROTOTYPES, my dataset has 69 classes. I use the default BG_PROTOTYPES: "weights/initial/background/background_prototypes.vitl14.pth", i get...
Hello! Thanks for your nice work. I am tring to run FSOD evaluation demo on coco dataset. But the inference speed of the inference phase is quite low on one...
