Yifei Pei
你就把他的code转成Tensorflow 2就行了。RTX30 系列的显卡需要更高版本的CUDA，而更高版本的CUDA不支持Tensorflow 1.8。
I also have this issue. But it's okay as long as the test images look good. I saw my test images look good.
I suggest you write your own dataloader and prepare a cropped image dataset so you don't need to crop images everytime.
Yes, writing own dataloader solve this issue.
Hi Ali, The predicted entropy is not real bits. Hyper-prior models might have low accuracy when predicting the real standard deviation if the bit rates are low. When the bit...
> Hi all, Sorry for the slow response. I think the theoretical bpp is computed as the entropy of the latents and hyperlatents as output by their respective probability models....
> Hi Yifei, thanks for your interest in HiFiC. I will look into this this week. could you provide some details? It is my understanding that HiFiC does in theory...
> Hi Yifei, thanks for your interest in HiFiC. I will look into this this week. could you provide some details? It is my understanding that HiFiC does in theory...
> Hi @yifeipet I successfully ported the LPIPS code from this repo to TF2.x by simply changing the outdated function calls to their respective TF2 counterparts, it works perfectly and...
I think the Pytorch MS SSIM is based on a differentiable MS SSIM algorithm so it's an approximation of the real MS SSIM.