Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
PMQ-VE: Progressive Multi-Frame Quantization for Video Enhancement
Authors: ZhanFeng Feng, Long Peng, Xin Di, Yong Guo, Wenbo Li, Yulun Zhang, Renjing Pei, Yang Wang, Yang Cao, Zheng-Jun Zha
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments demonstrate that our method outperforms existing approaches, achieving state-of-the-art performance across multiple tasks and benchmarks. The code will be made publicly available. |
| Researcher Affiliation | Academia | 1USTC, 2Max Planck Institute, 3CUHK, 4SJTU, 5Institute of Automation, Chinese Academy of Sciences, 6Chang an University EMAIL, EMAIL |
| Pseudocode | Yes | Algorithm 1 Backtracking-based Bound Initialization (BTBI) pipeline |
| Open Source Code | No | The code will be made publicly available. |
| Open Datasets | Yes | The Vimeo-90K [74] dataset is used for training across all tasks, with Vid4 [38] and the Vimeo-90K test set serving as evaluation benchmarks. |
| Dataset Splits | Yes | The Vimeo-90K dataset contains 64,612 training clips and 7,824 testing clips, with each clip consisting of seven frames at a spatial resolution of 448 256. For evaluation, in addition to the Vimeo-90K test set, we also adopt the Vid4 dataset [38], a classical benchmark containing four video sequences (calendar, city, foliage, and walk). ... The Vimeo-90K test set is further divided into Vimeo-Slow, Vimeo-Medium, and Vimeo-Fast subsets based on motion magnitude. |
| Hardware Specification | Yes | All experiments are implemented in Python with Py Torch [54] and conducted on 8 NVIDIA V100 GPUs. |
| Software Dependencies | No | All experiments are implemented in Python with Py Torch [54] and conducted on 8 NVIDIA V100 GPUs. |
| Experiment Setup | Yes | We adopt the Adam optimizer [27] with an initial learning rate of 2 10 4 and apply Cosine Annealing [45] over 20,000 iterations. The batch size is set to 8 and 2 per GPU during the initialization and distillation-based fine-tuning phases, respectively. |