Back to Search
Document details
AhG11: HOP training process and models
Abstract
This contribution gives information on the HOP training or fine-tuning process. In particular, information on models and results for partial retraining are given.
JVET-AE0289 AhG11: HOP training process and models [F. Galpin (InterDigital), S. Eadie, D. Rusanovskyy, Y. Li (Qualcomm), Y. Li, J. Li (ByteDance), L. Wang, R. Chang (Tencent), Z. Xie (Oppo), E. Alshina (Huawei)] [late]
This document provides detailed information about how to conduct the training when starting from stage 2 or stage 3. It shall be made available along with the model, and should be placed in the software repository as well.
There was no need for presentation of this contribution.
It was suggested that later more generic guidelines for training of NNVC models (not only HOP) could be developed from parts of this contribution.
Other aspects of neural network-based video coding (0)
Section kept as a template for future use.
AHG6/AHG12: Enhanced compression beyond VVC capability (78)
Summary and BoG reports
Contributions in this area were discussed at 1515–1815 on Tuesday 11 July 2023 and at 0830–1045 on Wednesday 12 July 2023 (chaired by JRO).
Decisions
Section kept as a template for future use.
Citation