在network.py的61行有什么特别的用意吗？

AndreGuo / ITMLUT

Official PyTorch implementation of "Redistributing the Precision and Content in 3D-LUT-based Inverse Tone-mapping for HDR/WCG Display" in CVMP2023 (SIGGRAPH European Conference on Visual Media Production).

Mozilla Public License 2.0

31 stars 3 forks source link

在network.py的61行有什么特别的用意吗？ #2

Closed zjysnow closed 5 months ago

zjysnow commented 5 months ago

如题 LUTs_d = self.lut_gen_b(weights_d) 这里为啥用lut_gen_b来生成LUTs_d？

AndreGuo commented 5 months ago

该行代码为笔误，我们马上修复该问题，修复完毕后将关闭该问题。 LUTs_d = self.lut_gen_b(weights_d) It's a mistake, we will fix it soon, it should be: LUTs_d = self.lut_gen_d(weights_d) We will fix it soon, and will close this issue once fixed.

AndreGuo commented 5 months ago

该问题已于network.py以及训练代码中修复。剩余模型参数尚需对应精调，新模型参数上传完毕后将关闭该问题。 This mistake is now fixed in network.py and the training code. Still need to refine the model parameters accordingly, we'll close this issue once new parameters are uplaoded.