Shuo Jin comments

Results 8 comments of


                                            Shuo Jin

关于BasicVSR++在Vimeo数据集训练的问题

> 你好，请问一下basicvsr++你是如何训练的？我没看到git上相关的引导以及训练细节，有没有相关的指令？请查阅mmedit教程文档：https://mmediting.readthedocs.io/zh_CN/latest/quick_run.html#id4 有详细指令解释

what is meaning of '0'

maybe it just means an object has been detected successfully. All objects are classfied as '0'

Error in basicsr/losses/losses.py

Maybe there is no defination of L2loss() or no registery, just add it yourself, not diffcult i think

关于自制训练数据集

maybe you need tensor(~~~, dtype=float)

Question about the pictures size of training and testing

while testing, the image in seq will be cut into patches and patches will be sent into the model and get the restored patches, then these patches will combined the...

请问一下如何用已有的模型，在自己的数据集上作推理，不用验证，是否有脚本文件

check files that named like "xxxx_inference.py"

Environment error when trying evaluation on F-ViT

This error is about MMDetection lib installation. Check the version between Torch, mmcv, Numpy, opencv might work for you. I further recommend using openmim for mmcv series installation.

Open Vocab Segmentation Not Working

Hi, about which resolution for the input image can obtain a good segmentation result?