Semantic-Segment-Anything
Semantic-Segment-Anything copied to clipboard
fudan-zvg
Automated dense category annotation engine that serves as the initial semantic labeling for the Segment Anything dataset (SA-1B).
It seems there are too many masks and labels for some simple images, is it possible to use dense crf to improve the masks/labels quality?
cuda:11.1 torch:1.10 single A6000 命令：python scripts/main_ssa_engine.py --data_dir=/mnt/usb/gxy/dataset_path/images --out_dir=output --world_size=1 --save_img --sam --checkpoint-path=checkpoint-path/sam_vit_h_4b8939.pth 报错：RuntimeError: CUDA out of memory. Tried to allocate 16.00 MiB (GPU 0; 47.54 GiB total capacity; 537.07 MiB...
Reference: https://github.com/ChaoningZhang/MobileSAM Our project performs on par with the original SAM and keeps exactly the same pipeline as the original SAM except for a change on the image encode, therefore,...
For the messy category generation data, such as "the word '50' in white letters," "three blue plastic rabbits", "three blue plastic snowflakes", "some very pretty blue and black items" and...
运行一会后报错
``` File "/media/admin1/envs/anaconda3/envs/leng_lip/lib/python3.7/site-packages/transformers/models/clipseg/modeling_clipseg.py", line 1458, in forward conditional_pixel_values=conditional_pixel_values, File "/media/admin1/envs/anaconda3/envs/leng_lip/lib/python3.7/site-packages/transformers/models/clipseg/modeling_clipseg.py", line 1360, in get_conditional_embeddings raise ValueError("Make sure to pass as many prompt texts as there are query images") ValueError: Make...
why the segformer result is lower than segformer paper display? is the different with evaluation metircs? i think the paper miou arrive to 84% , but the result you pointed...
使用非light-mode运行SSA-engine标注自己的数据集时时，不报错，也不输出结果，即使数据集里只有一张图片也这样，是因为GPU显存不足么，我的是rtx4060。使用light-mode模式运行时正常输出，会保存标注结果。
Hello, Thank you for your great work. I believe a lot of people can save time for their annotation works. Could you kindly share processed annotation data for the tested...
Hello all, I noticed that the semantic segmentation time with this tool can take around 1 minute per image. I was wondering if reducing the number of object categories to...
