SECourses Video and Image Upscaler Pro - STAR - SeedVR2 - Image Upscalers (e.g. 4xUltraSharp, RealESRGAN) - Industry Level Automatic Scene Based Movie Restoration

Patreon exclusive posts index to find our scripts easily, Patreon scripts updates history to see which updates arrived to which scripts and amazing Patreon special generative scripts list that you can use in any of your task.

Join discord to get help, chat, discuss and also tell me your discord username to get your special rank : SECourses Discord

Please also Star, Watch and Fork our Stable Diffusion & Generative AI  GitHub repository and join our Reddit subreddit and follow me on LinkedIn (my real profile)

=======

• Latest zip file Ultimate_Video_Processing_v7.zip

• Watch this video to see 4K upscale models comparison (it also has mini tutorial - public video) : https://youtu.be/q8QCtxrVK7g

• This app designed as an industry level video upscaling and restoration mindset, you will see as you use. Entire movies can be upscaled / restored with just 1 click fully automatically. Our image based upscalers supports full batch-size processing which means you can increase processing speed significantly.

• Our 1-click installer works on every GPU of RunPod and Massed Compute including RTX 5000 series and also on Windows same way

• App Screenshots Gallery : https://www.reddit.com/r/SECourses/comments/1m3ir9u/secourses_video_and_image_upscaler_pro_v1/

What is coming:

• Bug fixes if you find

• Fixing GPU ID selection for SeedVR2 image and video upscale

• More preset configs for SeedVR2 and more extensive testing

• SeedVR2 block swapping feature to significantly reduce VRAM

• SeedVR2 tiled-VAE - this will make massive difference

How To Use

• Generated videos will be saved inside STAR\outputs for all models and generated single upscale images will be saved inside STAR\outputs_images for all models

• It has a very detailed How to Use Documentation click to read after installing - on main page

  • Hopefully I will make a full detailed tutorial soon after SeedVR2 implementation

• Recomended upscaler is SeedVR2 - uses lesser VRAM and supports ratio upscale

• STAR model is working best at Max Frames per Chunk (VRAM) = 32

  • But its bigger limitation is it can only properly upscale 4x, if you do ratio upscale it is very low quality

  • 960p to 3840p uses around 145 GB VRAM with 32 frames, but 16 frames fits into 80 GB GPUs

  • You can also upscale 640p to 2560p

• SECourses Video and Image Upscaler Pro app has everything with automatic implementation just play a bit and you will see, you can upscale entire movie fully automatically

Windows Requirements

• Python 3.10.11, CUDA 12.8, cuDDN 8.9 and cuDNN 9.7+, FFmpeg, Git, C++ Build Tools 17.8 LTSC, MSVC

• Full tutorial to install above requirements : https://youtu.be/DrhUHnYfwC0

  • The above tutorial shows all the requirements step by step with the updated links in source post

15 August 2025 V8 Update

• Use latest zip file and run Windows_Install_and_Update.bat - should be sufficient

• A bug in image upscale fixed

• The dev repo has now full support of tiled VAE and block swap

• With both enabled, even low VRAM GPUs will be able to amazingly upscale videos and images with SeedVR2

  • Hopefully coming soon

31 July 2025 V7 Update

• Use latest zip file and run Windows_Install_and_Update.bat - should be sufficient

• Major SeedVR2 Improvements - Scene Split, RAM/VRAM Optimization & Enhanced Batch Processing

• Version 7.0 brings significant improvements to SeedVR2 upscaling with enhanced scene split processing, fixed memory leaks, improved single image pipeline, and comprehensive batch processing enhancements. This release focuses on stability, performance, and VRAM optimization for better user experience.

• Major Improvements: Scene split processing now fully working with SeedVR2, RAM/VRAM leakage fixed, single image pipeline perfected, and batch processing significantly enhanced.

• Core SeedVR2 Improvements

  • Scene Split Processing: Scene split based processing now fully working with SeedVR2 and is recommended for optimal results

  • Automatic Resolution Determination: Resolution detection and setting mechanism fixed for accurate upscaling

  • Single Image Pipeline: Single image processing pipeline fixed and working perfectly

  • Preset System: Preset save and load functionality fixed for SeedVR2 configurations

• Memory Optimization

  • RAM/VRAM Leak Fixed: Major memory leakage issues resolved with significantly improved VRAM usage patterns

  • VRAM Usage Optimization: 24 GB GPU can now upscale single images to 2560x2560px (6.55 megapixels) in under 30 seconds

  • Video VRAM Management: 24 GB GPU can upscale videos to 1280x720px with batch size 9 (VAE decoding uses 23.5 GB peak)

  • Batch Size Optimization: Due to model restrictions, batch size can only increase by 4 (5, 9, 13, 17, etc.)

• Batch Processing Enhancements

  • Batch Image Upscaling: Significantly improved with bug fixes and performance optimizations

  • Settings Synchronization: Batch image upscaling now uses settings from Single Image tab for consistency

  • Multi-Model Support: Batch upscale now works with both SeedVR2 and Image Upscalers like 2xLiveActionV1_SPAN_490000.pth

  • Quality Recommendation: SeedVR2 upscale recommended as it provides SUPIR-level quality or better depending on task

• Known Limitations

  • Block Swapping: Currently not working and temporarily disabled until tiled VAE encoding/decoding arrives

  • Tiled VAE Decoding: Not yet implemented but coming soon for significant VRAM reduction

  • FP8 Precision Models: Not useful yet pending further optimization

  • SeedVR2 3B FP16 Model: Has VRAM doubling issue - do not use until fixed

• Working Models & Recommendations

  • SeedVR2 7B FP16 Model: Working perfectly and recommended for best quality

  • Testing Presets: Start testing with SeedVR2_Image and SeedVR2_Video presets for optimal results

  • VRAM Management: If out of VRAM, reduce max resolution from Resolution & Scene Split settings

  • Performance Testing: RTX 3090 Ti (24GB) confirmed for all specified performance benchmarks

• Additional Fixes

  • Batch Size Slider: Fixed to properly increment by 4 as required by model architecture

  • Performance Improvements: Numerous other bug fixes and performance optimizations throughout the codebase

• Coming Soon: Tiled VAE implementation will significantly reduce VRAM usage, allowing larger upscale sizes with bigger batch sizes.

• Now supports MKV files too

  • Just run Windows_Install_and_Update.bat file and it will quickly update

• When uploading a video make selection filter type to all from right bottom of screen

Massed Compute (Recommend Cloud) :

• Please register via this link : https://vm.massedcompute.com/signup?linkId=lp_034338&sourceId=secourses&tenantId=massed-compute

  • We have a special coupon for all GPUs : SECourses

    If you want to learn more about GPUs and prices read this link : https://www.patreon.com/posts/126671823

    Select RTX A6000 or Better GPU - like L40S or A6000 ADA or A100 or H100

  • Then select our image SECourses from Creator dropdown

  • Then follow Massed_Compute_Instructions_READ.txt

  • Same as my any other Massed Compute installer script

  • Example tutorial for learn how to install and use Massed Compute

    • (Starts at 12:58) : https://youtu.be/KW-MHmoNcqo?si=G1WbG-Qw4ujWvOtG&t=778

RunPod (Cloud):

• Please register via this link : https://get.runpod.io/955rkuppqv4h

  • Then follow Runpod_Instructions_READ.txt

  • Same as my any other RunPod installer script

  • Use the template written in Runpod_Instructions_READ.txt file

  • Example tutorial for learn how to install and use RunPod

    • (starts at 22:03) : https://youtu.be/KW-MHmoNcqo?si=QN8X8Sjn13ZYu-EU&t=1323