dydxdt

> > In my opinion, in order to convert the pytorch model to onnx sucessfuly, the model should be rewriten. To get the right image result , this model should...

您好，我是从头训练了一遍REDS-width64的模型，使用3块V100（论文中8块），其他都保持不变（如400000个iteration）但是在REDS-val-300上得到的结果是28.83/0.862，和论文中29.09/0.867相差较大。请问您大概了解可能的原因吗

> @leonmakise 我们也提交了SIDD [online benchmark](https://www.eecs.yorku.ca/~kamel/sidd/benchmark.php) 的结果。 > > @dydxdt 您好， 为了达到文中所报的点， 训练设置(batch_siz应为64 / lr 1e-3 / iters 400k) 需要和原文对齐。 减少batch size其他保持一致， 则1）模型见到的数据少了， 2). 可能lr需要重新设置。 这两种情况都会导致效果不如预期。 谢谢您的回复！我再去试试

> Hi, meitounao110, Thanks for your attention to NAFNet. This speed is normal. But it is recommended to keep the overall batch size consistent with our config (to reproduce our...

I am working on the TF Lite model on Android, and I am getting into trouble for a while. If you understand it, can you help me with this: except...

> This repository is not for android, but you can ask @Jeongah-Shin who is an android developer on MoT Labs. Thank you very much!

Now I use tensoflow and try to find how to get the 'self.head.weight' in tensorflow. Or I have misunderstood and make it complex? Thank you for your reply.

Ok, I see. Thank you very much. Why don't you just use the normalization like batch norm or group norm? I'm not sure why you use the normalization method in...

> 958984240， 您好， > > Denoise的大图像测试和其他任务并无不同， 一般来说建议整图测试。 您好，我现在用自己的数据集，按原train.py代码在训练的时候进行val的测试，但是在测试部分图片后会报这样的错误 File "～/Projects/NAFNet/basicsr/train.py", line 305, in main() File "～/Projects/NAFNet/basicsr/train.py", line 270, in main model.validation(val_loader, current_iter, tb_logger, File "～/Projects/NAFNet/basicsr/models/base_model.py", line 55, in validation return...

Same question. 768 is inconsistent with the embed_dim(=[32, 64, 128, 512]) in the code. What's the relationship between them? I'm trying to train 3D medical images with this method and...