DeepMosaics
DeepMosaics copied to clipboard
HypoX64
Can't find pretrained models in archive
Hi,
I downloaded the Windows version of DeepMosaic to try out. Using the GUI, the inputs seem simple:
Input the target file (video) Input a pretrained model file
Using help, the list below shows the available pretrained files:
| Name | Description |
|---|---|
| add_face.pth | Add mosaic to all faces in images/videos. |
| clean_face_HD.pth | Clean mosaic to all faces in images/video.(RAM > 8GB). |
| add_youknow.pth | Add mosaic to all (FBI Warning) in images/videos. |
| clean_youknow_resnet_9blocks.pth | Clean mosaic to all (FBI Warning) in images/videos. |
| clean_youknow_video.pth | Clean mosaic to all (FBI Warning) in videos. |
| clean_youknow_video_HD.pth | Clean mosaic to all (FBI Warning) in videos.(RAM > 8GB) |
The issue is that these .pth files are not in the archive files in the program download.
Where do I get them?
Thanks,
Art
The pre-trained model is in the link I have shared.
https://drive.google.com/drive/folders/1LTERcN33McoiztYEwBxMuRjjgxh4DEPs
Hi, Thanks for your response. I am still not seeing the models. In the DeepMosiacs directory there is a folder called pretrained_models.:
When I open that folder there is only the message;
Should the models be one or both of the compressed files that the link points too. When I uncompressed those file, I do not see the models. What am I missing? Thanks, -Art
On Sunday, April 19, 2020, 8:01:29 AM PDT, Hypo <[email protected]> wrote:
The pre-trained model is in the link I have shared. https://drive.google.com/drive/folders/1LTERcN33McoiztYEwBxMuRjjgxh4DEPs
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.
Hi, I've figured out the pretrained files but here's what confusing- the Google screen that I see is this:
These are the original files that I downloaded and donot contain the training files. What you need to do is go to the upper right corner of the same page. You will see this:
Left click on the file icon and the two training files will be placed in .zip files. It takes some time but these zip files can can be unzipped and placed in the pretrained files directory in the DeepMosaics folder. When running the DeepMosaics binary, I presume that the path and training file desired (second selection) in the GUI prompt can be selected for processing. Haven't tried this yet on a mosaic video yet. Any tips? Thanks, -Art
On Sunday, April 19, 2020, 8:01:29 AM PDT, Hypo <[email protected]> wrote:
The pre-trained model is in the link I have shared. https://drive.google.com/drive/folders/1LTERcN33McoiztYEwBxMuRjjgxh4DEPs
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.
Oh! This current exe version is not support GPU. Processing videos will take a long time and I don't recommend to process video over 10 seconds.
This sounds bad. (boo how)
不好
So, what can I do? How to use most recent training files to remove mosaic for large video? > 1G
On Sunday, April 19, 2020, 8:26:50 PM PDT, Hypo <[email protected]> wrote:
Oh! This current exe version is not support GPU. Processing videos will take a long time and I don't recommend to process video over 10 seconds.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.
Am not sure how to do this. Why does not work with Windows GUI interface?
On Sunday, April 19, 2020, 9:03:54 PM PDT, Hypo [email protected] wrote:
Just like #7
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.
I try to build a exe with cuda, but failed . ( >﹏<。) I will fix this in the future.
when you can create it will be wonderful because you can't really do it on the cpu
Hi,One comment that I have though is that it runs really slow even on my workstation. It took a week to get through phase 1, locating the mosiacs. Now it's on phase 2 doing the mosaic cleanup which I estimate will take 3-4 days.
Still don't know what the result will be. Can the program code be optimized to run faster? Thanks, -Art
On Tuesday, April 28, 2020, 6:06:53 PM PDT, Sangenshin <[email protected]> wrote:
when you can create it will be wonderful because you can't really do it on the cpu
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.