hkchengrex/Tracking-Anything-with-DEVA
[ICCV 2023] Tracking Anything with Decoupled Video Segmentation
PythonNOASSERTION
Issues
- 1
mask problem
#101 opened by HibaDoi - 1
- 2
Floor tracking is poor. Any suggestion?
#98 opened by garrisonz - 1
How to Perform Inference on Multiple Videos?
#99 opened by weilllllls - 2
RuntimeError: shape '[1, 63, 512, 20, 27]' is invalid for input of size 19353600
#97 opened by GuanghaoJin - 29
something about data format
#85 opened by wkywwds - 3
Will changing the batch size from 16 to 18 and running with 6 cards result in performance loss?
#96 opened by longmalongma - 6
- 4
Can deva be extended to video semantic segmentation and video instance segmentation tasks?
#92 opened by longmalongma - 2
reading data is particularly slow
#94 opened by longmalongma - 1
Hello, I found that this method matches black results when there are only two images of this input.。。
#93 opened by chenkangjie1123 - 1
Subsampling BURST when evaluating
#91 opened by chhchan - 2
- 3
Something about the provided detection on VIPSeg
#88 opened by chhchan - 2
- 3
The mask in example
#87 opened by wkywwds - 7
- 1
extracting mask and bbox
#83 opened by shani1610 - 5
Just want to confirm!
#81 opened by JawadTawhidi - 2
yolo_world with DEVA
#78 opened by LLH-Harward - 1
Just want to confirm!
#82 opened by JawadTawhidi - 13
U-net with DEVA
#79 opened by wkywwds - 2
- 1
TrustImageSeg Inference
#77 opened by SangbeomLim - 4
0-5 represents the target?
#76 opened by longmalongma - 3
Didn't you use a mask that was inferred from a large model during semi-youtubevos inference?
#75 opened by longmalongma - 1
target tracking count
#74 opened by YuqiuDeng - 10
- 2
how to get the mask path to detections for rvos
#72 opened by xinbaiw - 3
- 6
- 2
Error in the middle of a video.
#67 opened by Mountchicken - 2
valueerror of text prompt
#70 opened by shangchengPKU - 1
text prompt
#69 opened by yangyang117 - 3
Deva out of memory
#54 opened by zhangzehao123 - 1
High Memory Usage During Initialization
#68 opened by kaankarakose - 1
- 1
Segmentation broken after 20 seconds or when total about more than 30 objects appeared in the video
#63 opened by huyen-do-helm - 1
Crop / segment the objects
#62 opened by tralala87 - 1
Deva on CPU
#61 opened by riccardo24-7 - 3
Is there a way to use DEVA with Cutie?
#58 opened by goneflash - 2
How to inference a video directly
#59 opened by wuyiqii - 1
- 1
Pre-trained model
#56 opened by cwt000297 - 2
None in category_ids
#57 opened by Bariona - 5
- 1
Deva run out of memory with "Including non-PyTorch memory, this process has 17179869184.00 GiB memory in use."
#55 opened by zhangzehao123 - 5
- 3
- 1
About training details
#50 opened by csustYyh