Karine-Huang / T2I-CompBench

[Neurips 2023] T2I-CompBench: A Comprehensive Benchmark for Open-world Compositional Text-to-image Generation

https://arxiv.org/pdf/2307.06350.pdf

Karine-Huang/T2I-CompBench Issues

Question about human-eval dataset
Updated 2 months ago5
Question about BLIPvqa code with reproduce scores of shape, texture
Closed 8 months ago2
CLIPScore implementation inconsistent with original paper?
Closed 4 months ago4
The problem in vqa_dataset.py
Updated 4 months ago1
Code for compared methods
Updated 5 months ago12
Details about the LoRA checkpoint released
Closed 6 months ago1
BLIP VQAeval default 8 noun phrases
Closed 7 months ago2
Question on reproduce the results
Closed 7 months ago6
Question about the parameter "np_num" in BLIP_vqa.py
Closed 7 months ago2
About the GLIP install
Updated 7 months ago6
Dense Captioning Model for Attribute Binding Eval?
Updated 7 months ago3
What is the number of pseudo images generated by Stable Diffusion for GORS fine-tuning?
Updated 7 months ago4
Question about scores reported in Table 1, Table2, and Table3 in the papers.
Closed 7 months ago1
Incorrect naming in dataset
Closed a year ago
Other potential detection models?
Updated a year ago2