Modelpublic

shreshthsaini/brightrate-lm-7b-multiexposure

sourceHugging Facemitupdated 20d agoView on Hugging Face

0likes16downloads

Model Card

BrightRate-LM 7B multi-exposure adapters

This PEFT adapter is one result from the BrightRate-LM controlled input and scaling study.

Base model

Qwen/Qwen2.5-VL-7B-Instruct

Input interface

Eight uniformly sampled frames are each rendered at -2, 0, and +2 stops. The 24 images are passed in temporal-major order.

Training data and recipe

Five adapters were trained independently on the five content-separated BrightVQ splits. Training uses two epochs, a three-epoch cosine schedule horizon, learning rate 1e-4, micro-batch 1, gradient accumulation 8, and rank-16 LoRA with alpha 32 and dropout 0.05. MOS targets are interpolated across five quality words. The root adapter is split 0; splits/split-1 through split-4 contain the remaining adapters.

Training data: BrightVQ.

Metrics

Held-out metrics for the five 420-video test splits:

Split	SROCC	PLCC	KRCC	RMSE
0	0.9110	0.9120	0.7347	5.7437
1	0.9311	0.9287	0.7666	5.0971
2	0.9175	0.9245	0.7469	5.0977
3	0.8904	0.8957	0.6996	5.9874
4	0.8760	0.8925	0.6928	5.7481
Mean	0.9052	0.9107	0.7281	5.5348

Intended use

This adapter is intended for research on no-reference perceptual quality assessment of user-generated HDR video. Scores are not calibrated for other datasets, display pipelines, or video domains.

Code and input construction are available in BrightRate-LM.

Citation

bibtex

@article{saini2026brightratelm,
  title   = {BrightRate-LM: Representation-Aware Quality Assessment for User-Generated HDR Video},
  author  = {Saini, Shreshth and Wang, Yilin and Birkbeck, Neil and Adsumilli, Balu and Bovik, Alan C.},
  journal = {Machine Vision and Applications},
  year    = {2026},
  note    = {Submitted}
}

Links

Code and evaluation: github.com/shreshthsaini/BrightRate-LM. Dataset: BrightVQ on Hugging Face. Related papers: Beyond8Bits, CVPR 2026 (arXiv 2603.00938) and CHUG, ICIP 2025 (arXiv 2510.09879).