Need RIFE 4.6 General Model .Where is RIFE 4.6 General Model

[dipu-sr]

I am looking for Rife 4.6 general model .But I can't find it. i need Rife 4.6 general model. many anime editors said that 4.6 is the best model right now. that's why i need RIFE 4.6 model and i am going to use in google colab by replacing your rife practical notebook 4.1 model i hope that you giving me this model which is work in google colab . Thank you

[RazFerroo]

Huh, I can't find a download link for it either. Because of that, I've uploaded a Google Drive .zip containing the following:

• folder ''rife-cuda'' containing folders & files needed for RIFE 4.6 CUDA version; if I'm not mistaken, CUDA models have the fastest processing but are only available for NVIDIA GPUs.

• folder ''rife-ncnn-vs'' containing folders & files needed for RIFE 4.6 Vulkan NCNN/VS version; this is compatible with all GPUs, might be slightly slower, unsure.

These files are taken from my Flowframes installation, so the directory structure may not be the same as what your programs require. You'll need to shuffle/rename files as needed for whatever program you use.

If you are not using Flowframes, you need to edit a line in the IFNet_HDv3.py file for RIFE 4.6 CUDA:
(If you are using Flowframes, you can ignore this section.)
(Also, you don't need to do this for the rife-ncnn-vs version.)

1. From the linked .zip above, extract the contents anywhere you like.
2. Navigate to the folder [...]\pkgs\rife-cuda\RIFE46\arch.
3. In the arch folder, use a text editor to open the file IFNet_HDv3.py.
4. Find this line: from arch.IFNet_HDv3 import *
5. ... and replace that line with this:
   from train_log.IFNet_HDv3 import *
6. Save & Close the file.

Oh! And if you're using Flowframes, make sure to extract the contents of the .zip to [...]\Flowframes\FlowframesData\pkgs. If you are asked to overwrite any files, do not overwrite. It is preferable to keep what is already there to mitigate the risk of breaking the Flowframes installation.

I have no experience with Google Collab, so if you run into issues with that, you may need to seek further assistance elsewhere.

――――――――――――――――――――――――――――――――――――――――

As an aside, newer versions of RIFE are a lot better in my experience. RIFE 4.17 and RIFE 4.26 are the best of the best. I'd recommend using those over 4.6, but you can test it yourself to see the difference. Both of those models are available from links on the Practical-RIFE Github README.

――――――――――――――――――――――――――――――――――――――――

[dipu-sr]

Huh, I can't find a download link for it either. Because of that, I've uploaded a Google Drive .zip containing the following:

• folder ''rife-cuda'' containing folders & files needed for RIFE 4.6 CUDA version; if I'm not mistaken, CUDA models have the fastest processing but are only available for NVIDIA GPUs.
• folder ''rife-ncnn-vs'' containing folders & files needed for RIFE 4.6 Vulkan NCNN/VS version; this is compatible with all GPUs, might be slightly slower, unsure.

These files are taken from my Flowframes installation, so the directory structure may not be the same as what your programs require. You'll need to shuffle/rename files as needed for whatever program you use.

If you are not using Flowframes, you need to edit a line in the IFNet_HDv3.py file for RIFE 4.6 CUDA: (If you are using Flowframes, you can ignore this section.) (Also, you don't need to do this for the rife-ncnn-vs version.)

1. From the linked .zip above, extract the contents anywhere you like.
2. Navigate to the folder [...]\pkgs\rife-cuda\RIFE46\arch.
3. In the arch folder, use a text editor to open the file IFNet_HDv3.py.
4. Find this line: from arch.IFNet_HDv3 import *
5. ... and replace that line with this:
   from train_log.IFNet_HDv3 import *
6. Save & Close the file.

Oh! And if you're using Flowframes, make sure to extract the contents of the .zip to [...]\Flowframes\FlowframesData\pkgs. If you are asked to overwrite any files, do not overwrite. It is preferable to keep what is already there to mitigate the risk of breaking the Flowframes installation.

I have no experience with Google Collab, so if you run into issues with that, you may need to seek further assistance elsewhere.

――――――――――――――――――――――――――――――――――――――――

As an aside, newer versions of RIFE are a lot better in my experience. RIFE 4.17 and RIFE 4.26 are the best of the best. I'd recommend using those over 4.6, but you can test it yourself to see the difference. Both of those models are available from links on the Practical-RIFE Github README.

――――――――――――――――――――――――――――――――――――――――

tnx bro