gpt4vision/OvSGTR
[ECCV 2024 Best Paper Candidate] Implementation of "Expanding Scene Graph Boundaries: Fully Open-vocabulary Scene Graph Generation via Visual-Concept Alignment and Retention"
PythonNOASSERTION
Issues
- 2
Query about evaluation
#22 opened by focus1224 - 2
Results in Table 5
#21 opened by qwerhk839 - 0
Inquiry about Scene Graph Generation Code
#20 opened by bddk520 - 2
Quick Question : why not normalize the output of relation feature ? (OVR or OVD + R setting)
#19 opened by LimGeunTaekk - 8
Query about the pre-training process
#14 opened by qwerhk839 - 2
evaluation on custom datasets
#18 opened by Guapicat0 - 1
Generating the graphs
#17 opened by carloscaetano - 8
Training issues
#16 opened by ZMH-SDUST - 0
Unable to run knowledge distillation!
#15 opened by ZMH-SDUST - 1
RROR: modelname:groundingdino not in models:{}
#12 opened by ZMH-SDUST - 6
About the pretrain command on the COCO datset
#13 opened by myukzzz - 5
Improve HF integration
#11 opened by NielsRogge - 1
Requires files in the dataset
#10 opened by thebigeagle - 2
- 0
ModuleNotFoundError with GroundingDINO during Inference (Fixed by pip install -e)
#8 opened by zhuhl0913 - 1
Clarification on the Generation of image_data.json, VG-SGG-dicts.json, and VG-SGG.h5 for VG150
#7 opened by zhuhl0913 - 1
- 2
ECCV 2024 cam-ready version
#5 opened by LimGeunTaekk - 1
Can not run your code successfully
#3 opened by hnczgytmp - 1
Problem with config file
#4 opened by SIST-wangchongyu - 2
Hello, can you provide the vg_stats.pt # frequency bias in the data ├─ zeroshot_triplet.pytorch and ├─ VG-SGG-dicts.json # add split_GLIPunseen ├─ VG-SGG.h5
#2 opened by CarlBruce - 1
Code release
#1 opened by SanMumumu