[Feature]: optional upscale when face not found

[brknsoul]

First, confirm

• I have read the instruction carefully
• I have searched the existing issues
• I have updated the extension to the latest version

What happened?

When batching images (eg: extracted frames from a video), if a face isn't found (for example, the subject in that particular frame is turned away), the upscaler won't happen on that image.

Steps to reproduce the problem

Img2Img batch a folder of files with one or more images without a face.
Use upscaler.

Sysinfo

Python 3.10.11 on Windows
Version: app=sd.next updated=2023-09-28 hash=ccae1ef0 url=https://github.com/vladmandic/automatic/tree/master
Latest published version: 01c466632d299c087a9033f1846558e94383d132 2023-10-01T01:08:55Z
Platform: arch=AMD64 cpu=Intel64 Family 6 Model 158 Stepping 11, GenuineIntel system=Windows release=Windows-10-10.0.19045-SP0
python=3.10.11
Using DirectML Backend

Relevant console log

*Successful detection and upscaling*

14:56:47-958701 DEBUG    Sampler: sampler=DPM++ 2M Karras config={'scheduler': 'karras'}                                                                       
100%|##########| 1/1 [00:01<00:00,  1.62s/it]
14:56:38-020571 STATUS   Working: source face index [0], target face index [0]                                                                                 
14:56:38-027597 STATUS   Detecting Source Face, Index = 0                                                                                                      
14:56:39-781531 STATUS   Detected: -33- y.o. Female                                                                                                            
14:56:39-783528 STATUS   Detecting Target Face, Index = 0                                                                                                      
14:56:41-548553 STATUS   Detected: -26- y.o. Female                                                                                                            
14:56:41-550555 STATUS   Swapping Source into Target                                                                                                           
14:56:42-795001 STATUS   Restoring the face with CodeFormer                                                                                                    
14:56:43-351562 STATUS   Upscaling with 4x_NMKD-Siax_200k scale = 2                                                                                            
14:56:43-431571 STATUS   Model loaded: type=ESRGAN model=E:\SM\Data\Packages\SD.Next Web UI\models\ESRGAN\4x_NMKD-Siax_200k.pth                                
Upscaling ---------------------------------------- 100% 0:00:00 0:00:01
14:56:45-739760 STATUS   Processed: images=1 time=12.51s its=1.20 memory={'ram': {'used': 5.99, 'total': 15.95}, 'gpu': {'used': 7.54, 'total': 7.98},         
                         'retries': 'DirectMLDevice', 'oom': 0}                                                                                                
14:56:46-003635 DEBUG    Saving: image="D:\zvidout\date0014-0000-20231001145646.png" type=PNG size=1024x768                                                    
14:56:47-921837 DEBUG    Processed: images=20 memory={'ram': {'used': 5.98, 'total': 15.95}, 'gpu': {'used': 7.54, 'total': 7.98}, 'retries': 'DirectMLDevice',
                         'oom': 0} op=batch                                                                                                                    


*Unsuccessful detection, no upscaling*

14:56:47-958701 DEBUG    Sampler: sampler=DPM++ 2M Karras config={'scheduler': 'karras'}                                                                       
100%|##########| 1/1 [00:01<00:00,  1.62s/it]
14:56:52-639873 STATUS   Working: source face index [0], target face index [0]                                                                                 
14:56:52-647108 STATUS   Detecting Source Face, Index = 0                                                                                                      
14:56:54-381563 STATUS   Detected: -33- y.o. Female                                                                                                            
14:56:54-383535 STATUS   Detecting Target Face, Index = 0                                                                                                      
14:56:55-718400 ERROR    Age Detection: No face with index = 0 was found                                                                                       
14:56:55-720400 ERROR    Gender Detection: No face with index = 0 was found                                                                                    
14:56:57-018473 ERROR    Age Detection: No face with index = 0 was found                                                                                       
14:56:57-020473 ERROR    Gender Detection: No face with index = 0 was found                                                                                    
14:56:57-128989 STATUS   No target face found for 0                                                                                                            
14:56:57-132958 STATUS   Processed: images=1 time=9.17s its=1.64 memory={'ram': {'used': 5.98, 'total': 15.95}, 'gpu': {'used': 7.54, 'total': 7.98},          
                         'retries': 'DirectMLDevice', 'oom': 0}                                                                                                
14:56:57-203027 DEBUG    Saving: image="D:\zvidout\date0015-0000-20231001145657.png" type=PNG size=512x384                                                     
14:56:57-738819 DEBUG    Processed: images=20 memory={'ram': {'used': 5.98, 'total': 15.95}, 'gpu': {'used': 7.54, 'total': 7.98}, 'retries': 'DirectMLDevice',
                         'oom': 0} op=batch

Additional information

Essentially, if one is batching images, one would like them to be all the same size, regardless of face detection or not.

[Gourieff]

Ok, I will add it as an option in future updates