Home Creators Posts Import Register Favorites Logout

Downloads

Content

SECourses Video and Image Upscaler Pro restores entire movies automatically with scene-based STAR and SeedVR2 upscaling plus batch image upscalers.

Note: this is an older post kept for reference. The current, maintained version is SECourses Upscaler Pro: SeedVR2 & FlashVSR+ Video and Image Upscaler.

Industry level, automatic scene based movie restoration in 1 click, with image upscalers such as 4xUltraSharp and RealESRGAN and full batch-size processing. 1-click installers for Windows, RunPod and Massed Compute; V7 fixed RAM/VRAM leaks, and a 24 GB GPU upscales an image to 2560x2560 in under 30 seconds.

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)

=======

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) :

RunPod (Cloud):

Related posts

All posts in this topic: Image and Video Upscaling and Restoration · New to SECourses? Start Here: Best and Most Popular SECourses Posts

Files

SECourses Video and Image Upscaler Pro STAR vs TOPAZ StarLight vs Image Based Best Upscalers

This is a demo video of our SECourses Video and Image Upscaler Pro app. Hopefully SeedVR2 coming soon with ultra optimized Video and Image upscaling feature. App here : https://www.patreon.com/posts/134405610 Also hopefully I will make a ultra detailed tutorial video too soon. 🔗Follow below link to download the zip file that contains SECourses Video and Image Upscaler Pro App ⤵️ ▶️ https://www.patreon.com/posts/SECourses-Video-and-Image-Upscaler-Pro-134405610 🔗Follow below link to download the zip file that contains Image and Video Comparison Slider App⤵️ ▶️ https://www.patreon.com/posts/Image-Video-Comparison-Slider-App-133935178 🔗 Python, Git, CUDA, C++, FFMPEG, MSVC installation tutorial - needed for ComfyUI ⤵️ ▶️ https://youtu.be/DrhUHnYfwC0 🔗 SECourses Official Discord 10500+ Members ⤵️ ▶️ https://discord.com/servers/software-engineering-courses-secourses-772774097734074388 🔗 Stable Diffusion, FLUX, Generative AI Tutorials and Resources GitHub ⤵️ ▶️ https://github.com/FurkanGozukara/Stable-Diffusion 🔗 SECourses Official Reddit - Stay Subscribed To Learn All The News and More ⤵️ ▶️ https://www.reddit.com/r/SECourses/ Some background music by NoCopyrightSounds : https://gist.github.com/FurkanGozukara/681667e5d7051b073f2e795794c46170

Comments

AiInfluence

It's finally here! Thank you Dr for your dedication and passion on designing this amazing app.

SECourses

you are welcome thank you as well

Neo Geo

Hey doc, any plans to add support for other video formats like mkv, avi? Thanks!

Unrecognised

so not usable with 16gb vram?

Neo Geo

Input, when I browse I can only see mp4s for selection?

Unrecognised

are any of them superior to gigapixel?

Sirens Talent Agency

any chance that this will work with an amd MI300x on runpod?

Lou

Thanks for the release. Have been looking forward to this for some time. I'll see if I can get some hands on testing during the next week or 2. Thanks again.

SECourses

awesome. hopefully SeedVR2 coming this week

SECourses

it works on B200 but I didn't test on AMD so can't tell

SECourses

i would say compare. i liked ultrasharp2

SECourses

hello now supports. gradio will show you not playable error but i just upscaled. upgrade to v2 . also click to show all types : https://ibb.co/jZBNLF3x

Ori Naor

Hi thanks for the amazing content, I have an error on batch upsclaing: Traceback (most recent call last): File "H:\repo\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\gradio\queueing.py", line 626, in process_events response = await route_utils.call_process_api( File "H:\repo\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\gradio\route_utils.py", line 350, in call_process_api output = await app.get_blocks().process_api( File "H:\repo\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\gradio\blocks.py", line 2235, in process_api result = await self.call_function( File "H:\repo\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\gradio\blocks.py", line 1746, in call_function prediction = await anyio.to_thread.run_sync( # type: ignore File "H:\repo\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\anyio\to_thread.py", line 56, in run_sync return await get_async_backend().run_sync_in_worker_thread( File "H:\repo\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\anyio\_backends\_asyncio.py", line 2470, in run_sync_in_worker_thread return await future File "H:\repo\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\anyio\_backends\_asyncio.py", line 967, in run result = context.run(func, *args) File "H:\repo\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\gradio\utils.py", line 917, in wrapper response = f(*args, **kwargs) File "H:\repo\SECourses_Upscaler_Pro_v1\STAR\secourses_app.py", line 3487, in batch_wrapper result =process_batch_videos_wrapper (build_batch_app_config_from_ui (*args )) File "H:\repo\SECourses_Upscaler_Pro_v1\STAR\secourses_app.py", line 3293, in build_batch_app_config_from_ui ( ValueError: too many values to unpack (expected 97)

고미 다이

When running Windows_Install_and_Update, it will be installed and then stopped. Why is that? If you forcefully shut down, you will get an error message .gitattributes: 1.52kB [00:00, ?B/s] | 0/7 [00:00 snapshot_download( File "C:\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\huggingface_hub\utils\_validators.py", line 114, in _inner_fn return fn(*args, **kwargs) File "C:\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\huggingface_hub\_snapshot_download.py", line 327, in snapshot_download thread_map( File "C:\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\tqdm\contrib\concurrent.py", line 69, in thread_map return _executor_map(ThreadPoolExecutor, fn, *iterables, **tqdm_kwargs) File "C:\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\tqdm\contrib\concurrent.py", line 51, in _executor_map return list(tqdm_class(ex.map(fn, *iterables, chunksize=chunksize), **kwargs)) File "C:\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\tqdm\std.py", line 1181, in __iter__ for obj in iterable: File "C:\Users\User\AppData\Local\Programs\Python\Python310\lib\concurrent\futures\_base.py", line 621, in result_iterator yield _result_or_cancel(fs.pop()) File "C:\Users\User\AppData\Local\Programs\Python\Python310\lib\concurrent\futures\_base.py", line 319, in _result_or_cancel return fut.result(timeout) File "C:\Users\User\AppData\Local\Programs\Python\Python310\lib\concurrent\futures\_base.py", line 458, in result return self.__get_result() File "C:\Users\User\AppData\Local\Programs\Python\Python310\lib\concurrent\futures\_base.py", line 403, in __get_result raise self._exception File "C:\Users\User\AppData\Local\Programs\Python\Python310\lib\concurrent\futures\thread.py", line 58, in run result = self.fn(*self.args, **self.kwargs) File "C:\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\huggingface_hub\_snapshot_download.py", line 301, in _inner_hf_hub_download return hf_hub_download( File "C:\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\huggingface_hub\utils\_validators.py", line 114, in _inner_fn return fn(*args, **kwargs) File "C:\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\huggingface_hub\file_download.py", line 988, in hf_hub_download return _hf_hub_download_to_local_dir( File "C:\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\huggingface_hub\file_download.py", line 1290, in _hf_hub_download_to_local_dir _download_to_tmp_and_move( File "C:\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\huggingface_hub\file_download.py", line 1710, in _download_to_tmp_and_move xet_get( File "C:\SECourses_Upscaler_Pro_v1\venv\lib\site-packages\huggingface_hub\file_download.py", line 627, in xet_get download_files( KeyboardInterrupt ^C일괄 작업을 끝내시겠습니까 (Y/N)? Even if I wait for hours, the download has stopped. What should I do? ㅜㅜ

SECourses

i recommend install warp vpn and run installer while it is on. it will 99% fix your issue. warp vpn is from cloudflare 100% secure

Anshul Gupta

how to enhance image and how much VRAM is required to upscale ideo? I have an RTX 40 4090

고미 다이

gitattributes: 1.52kB [00:00, ?B/s] | 0/7 [00:00

SECourses

for image enhance you can use image based upscalers. for 4090 currently you can use image based upscalers in core settings. for STAR model you need at least 80 GB GPUs. I am working on SeedVR2 for lower vRAM

SECourses

thank you for reporting fixed with V3. Just run Windows_Install_and_Update.bat file and it will quickly update

SECourses

please install Warp VPN and run with it. it is from cloudflare so 100% secure. it should fix your download issue

Anshul Gupta

but how to upload image there is no image uploading option

SECourses

you are right. it is coming with SeedVR2. currently you can process images as folder. in batch upscaling use "Direct Image Upscaling" checkbox

고미 다이

Sir, even with Warp VPN, it's still pausing during installation. I feel blue.

SECourses

well it is related to your computer. something blocking your installation you have to figure out. it can be antivirus as well

SECourses

if you upgrade gold member i can connect your pc and try to figure out what is blocking your access to github and hugging face

Juanmyth

Standalone face restoration seems to only detect/restore a few frames (5 out of 162 frames all with the same face) and produce a very short video.

SECourses

i just tested and it worked on entire video. it was even a hard video. is face tiny hard to detect etc? but i see final frame count is not accurate. it is lesser. i am working on a fix now please wait

SECourses

thank you for reporting. With V5 Single face restoration was broken Missing frames, inaccurate FPS and more and all fixed It will also save extracted frames and processed frames as well into outputs like inside 0005 folder Final restored video will be also saved inside such folder like 0005

cool1

The demo looks good and it does look higher resolution and sharper, but maybe the upscaling could be made a bit more natural. eg. in the teeth and beard it looks a bit CGI-like (though the teeth may all be caps or veneers). Maybe it depends what it's trained and whether it has enough human faces in the (probably video) training data. I assume the model is given a low res and a high res source it needs to learn to generate.

Hoàng Giang Sơn Trương

help me Traceback (most recent call last): File "F:\AI\SECourses_Upscaler_Pro_v1\STAR\logic\upscaling_core.py", line 496, in run_upscale util_run_ffmpeg_command (cmd ,"Input Downscaling with Audio Copy",logger =logger ) File "F:\AI\SECourses_Upscaler_Pro_v1\STAR\logic\ffmpeg_utils.py", line 83, in run_ffmpeg_command raise gr.Error(f"ffmpeg command failed (see console for details): {e.stderr.strip()[:500] if e.stderr else 'Unknown ffmpeg error'}") gradio.exceptions.Error: "ffmpeg command failed (see console for details): 'ffmpeg' is not recognized as an internal or external command,\noperable program or batch file."

고미 다이

When will seedVr2 be released? I'm so curious.

cjpo

Will there be support for multi GPU? I am sure at the moment there is not and you can only select one GPU from the drop down box?

SECourses

saldy no multi GPU yet. i just added SeedVR2 it is way more amazing than STAR at the moment

SECourses

i just added SeedVR2 it is way more amazing than STAR at the moment

고미 다이

If I go through the Windows_Download_Models_and_or_Resume or Windows_Install_and_Update bat file, can I use SeedVR2? I tried to sell them both and no video scaler, and they said there is no SEEDVR2 MODEL. Do I need to download the seedvr2 3B model separately and put it in the folder?

Glen Carpenter

May want to include a note about the disk space requirements as well, I think everything came out to around 87 GB

Realtoonai

Is this working for rtx 4090 24GB ? How much hard disk capacity needed?

SECourses

yes there are models you can use for 24 gb. it downloads a lot of models and you can delete the ones you dont need.

SECourses

true it downloads a lot of models but you can delete the ones you dont need. i will make models explanation

SECourses

it downloads seedvr2 3b and 7b models into SeedVR2\models automatically

Juanmyth

In new version (v5) I get this error instead: "Processing failed: Processing failed due to unexpected error: too many values to unpack (expected 2)." The tested videos were 5 seconds, 32 fps or 16 fps with different resolutions (for instance 832x1088, 512x768). Nothing big. I did a fresh installation of V5.

고미 다이

Running Windows_Install_and_Update or Windows_Download_Models_and_or_Resume will not download the model. Singing Windows_Download_Models_and_or_Resume distribution date *Running on local URL: http://127.0.0.1:7860 * To create a public link, set `share=True` in `launch()`. 13:39:11 - INFO - SeedVR2 models directory extensions but no models found, is there a solution? I'm using Cloudflare WARP even if I delete and reinstall all files just in case.

SECourses

well sadly something blocking downloads in that case. i am not sure. you can try to download manually. i may add manual download link to top soon

SECourses

hi where do you get this error?

Juanmyth

In Face Restoration.

ApathyJones

Really nice work with this! Lots of really advanced features, especially in the GUI. Simple things like being able to zoom and scroll images and stuff in the comparison windows is great!

고미 다이

Download successful. I entered the Google DNS address without Warp and the download was completed. ^^ The graphics card on my PC is Geforce 4090 / 24GB. When I upscaled a 34-second 1920x1080 video using a SeedVR2 model with a downscale then upgrade, the process ran for about 34 minutes, but the video came out black. Is there anything I can check? I tried 10 times, so it's the same.

SECourses

pleas wait v7 update it will be fixing so many stuff and upscale too. with 24 gb you can upscale to 2560x2560 on empty gpu - seedvr2 image upscale

SECourses

thanks and many more coming. i did huge improvements with v7 about to publish tomorrow

JAV Pomoe

Hello, Dr. I uploaded the original video and when I pressed the Upscale Video button, I got the following error. I was using V7. Is there a solution? Hugging Face cache folders set to: D:\Ultimate_Video_Processing_v7\STAR\models VRAM libre: 22.46 GB Using STAR repository base_path: D:\Ultimate_Video_Processing_v7\STAR W0803 17:26:02.207000 29672 Lib\site-packages\torch\distributed\elastic\multiprocessing\redirects.py:29] NOTE: Redirects are currently not supported in Windows or MacOs. 17:26:03 - INFO - Diagnostic: Explicitly set StreamHandler level to INFO. 17:26:03 - INFO - Logger 'video_to_video' configured with level: INFO. Handlers: [ (INFO)>] 17:26:03 - INFO - Found last used preset: 'STAR'. Attempting to load. 17:26:03 - INFO - Successfully loaded initial preset 'STAR'. 17:26:03 - INFO - Setting seedvr2.model from preset: 'seedvr2_ema_3b_fp16.safetensors' 17:26:03 - INFO - Initial preset for dropdown: 'STAR' (actually loaded) 17:26:04 - INFO - Scanning for image upscaler models in: D:\Ultimate_Video_Processing_v7\STAR\upscale_models 17:26:04 - INFO - Found 13 image upscaler model(s): ['2x-AnimeSharpV4_Fast_RCAN_PU.safetensors', '2xLiveActionV1_SPAN_490000.pth', '2xNomosUni_span_multijpg_ldl.pth', '2x_AniScale2_Omni_i16_40K.pth', '4x-AnimeSharp.safetensors', '4x-UltraSharp.safetensors', '4x-UltraSharpV2.safetensors', '4xNomos2_hq_dat2.safetensors', '4xRealWebPhoto_v4_dat2.safetensors', 'HAT-L_SRx4_ImageNet-pretrain.safetensors', 'Kim2091-4x-UltraSharp.safetensors', 'RealESRGAN_x4plus.safetensors', 'RealESRGAN_x4plus_anime_6B.safetensors'] 17:26:04 - INFO - [INIT] Converting SeedVR2 filename 'seedvr2_ema_3b_fp16.safetensors' to display name 17:26:04 - INFO - [INIT] Looking for filename 'seedvr2_ema_3b_fp16.safetensors' in 3 available models 17:26:04 - INFO - [INIT] Found exact match: seedvr2_ema_3b_fp16.safetensors -> SeedVR2 3B (FP16) 17:26:04 - INFO - SeedVR2 Model Dropdown - Config model: seedvr2_ema_3b_fp16.safetensors, Choices: ['SeedVR2 3B (FP16)', 'SeedVR2 7B (FP16)', 'SeedVR2 7B (FP16) - Enhanced sharpness variant'], Selected: SeedVR2 3B (FP16) 17:26:04 - INFO - Gradio App Starting. Default output to: D:\Ultimate_Video_Processing_v7\STAR\outputs 17:26:04 - INFO - STAR Models expected at: D:\Ultimate_Video_Processing_v7\STAR\pretrained_weight\light_deg.pt, D:\Ultimate_Video_Processing_v7\STAR\pretrained_weight\heavy_deg.pt 17:26:04 - INFO - CogVLM2 Model expected at: D:\Ultimate_Video_Processing_v7\STAR\models\CogVLMFixed 17:26:04 - INFO - Initialized with default GPU: GPU 0: NVIDIA GeForce RTX 4090 (24.0GB) (GPU 0) 17:26:04 - INFO - Effective Gradio allowed_paths: ['C:\\', 'D:\\', 'D:\\Ultimate_Video_Processing_v7\\STAR', 'D:\\Ultimate_Video_Processing_v7\\STAR\\logic', 'E:\\', 'F:\\', 'X:\\', 'Y:\\', 'Z:\\'] * Running on local URL: http://127.0.0.1:7860 * To create a public link, set `share=True` in `launch()`. 17:26:06 - INFO - Initializing SeedVR2 tab, keeping existing model selection 17:26:17 - INFO - Video info (fast) for 'winking_01.mp4': 41 frames, 25.00 FPS, 1.640s, 720x480, "mov 17:26:17 - INFO - Video info (fast) for 'winking_01.mp4': 41 frames, 25.00 FPS, 1.640s, 720x480, "mov 17:26:17 - INFO - Video uploaded: winking_01.mp4 17:26:17 - INFO - Video details: 41 frames, 25.00 FPS, 1.64s, 720x480 17:26:17 - INFO - Video info (fast) for 'winking_01.mp4': 41 frames, 25.00 FPS, 1.640s, 720x480, "mov 17:26:17 - INFO - Auto-resolution calculation: 720x480 → 1280x854 17:26:17 - INFO - Aspect ratio: 1.500 → 1.499 (error: 0.08%) 17:26:17 - INFO - Constraint usage: 1,093,120 / 1,638,400 (66.7%) 17:26:17 - INFO - Auto-resolution: C:\Users\Masato\AppData\Local\Temp\gradio\2ff999678a6130450139bcee62172b3ed4124954a8e0d2f0dd1f9628f956203d\winking_01.mp4 -> 1280x854 (was 1280x1280) 17:26:17 - INFO - Auto-resolution status: ✅ Auto-calculated: 1280x854 (1,093,120 pixels, 66.7% of constraint, aspect ratio error: 0.08%) 17:26:17 - INFO - Video info (fast) for 'winking_01.mp4': 41 frames, 25.00 FPS, 1.640s, 720x480, "mov 17:26:17 - ERROR - Error calculating expected output resolution: name 'app_config' is not defined 17:26:17 - INFO - Video info (fast) for 'winking_01.mp4': 41 frames, 25.00 FPS, 1.640s, 720x480, "mov 17:26:17 - INFO - Downscaling required: 720x480 -> 320x214 for 4.0x STAR target. 17:26:17 - INFO - Direct 4.0x upscale would exceed constraint box. Using optimal resolution within constraint. 17:26:17 - INFO - Target resolution mode: Downscale then 4.0x. Calculated upscale: 4.00x. Target output: 1280x854 17:26:17 - INFO - Video info (fast) for 'winking_01.mp4': 41 frames, 25.00 FPS, 1.640s, 720x480, "mov 17:26:17 - INFO - Pixel budget: 1,093,120 pixels, Final output: 1,093,120 pixels (100.0% of budget) 17:26:17 - INFO - Aspect ratio preserved: 1.500 -> 1.499 17:26:17 - INFO - Calculated final target resolution: 1280x854 with upscale 4.00 Traceback (most recent call last): File "D:\Ultimate_Video_Processing_v7\venv\lib\site-packages\gradio\queueing.py", line 626, in process_events response = await route_utils.call_process_api( File "D:\Ultimate_Video_Processing_v7\venv\lib\site-packages\gradio\route_utils.py", line 350, in call_process_api output = await app.get_blocks().process_api( File "D:\Ultimate_Video_Processing_v7\venv\lib\site-packages\gradio\blocks.py", line 2235, in process_api inputs = await self.preprocess_data( File "D:\Ultimate_Video_Processing_v7\venv\lib\site-packages\gradio\blocks.py", line 1897, in preprocess_data processed_input.append(block.preprocess(inputs_cached)) File "D:\Ultimate_Video_Processing_v7\venv\lib\site-packages\gradio\components\slider.py", line 154, in preprocess Number.raise_if_out_of_bounds(payload, self.minimum, self.maximum) File "D:\Ultimate_Video_Processing_v7\venv\lib\site-packages\gradio\components\number.py", line 128, in raise_if_out_of_bounds raise Error(f"Value {num} is less than minimum value {minimum}.") gradio.exceptions.Error: 'Value 1 is less than minimum value 5.'

SECourses

minimum batch size has to be 5 for seedvr

Nicolas Giarrusso

Hi! I started trying to use the simple still image upscaler, I set seedvr2 and when I hit run it showed me the error Value 1 is less than minimum value 5.

SECourses

ok it is like this. select seedvr image preset, go single image upscale, select image and click image upscale . it works. also make sure to run Windows_Install_and_Update.bat

Jon Sheehan

Hello! It's giving me an error that there no SeedVR2 models. In the SeedVR2/models folder I see it has downloaded: seedvr2_ema_7b_sharp_fp16.safetensors, seedvr2_ema_7b_fp16.safetensors, seedvr2_ema_3b_fp16.safetensors and ema_vae_fp16.safetensors

Jon Sheehan

Have run out of ideas of how to fix...

SECourses

please email me full error logs and screenshot of your folders : monstermmorpg@gmail.com

SECourses

also do not run anything as administrator

Jon Sheehan

Have emailed, thank you!

Alex

I add a video and enable the SeedVR2_Video enhancement, it processes up to 80%, then it stops, then it appears again and the right video screen is black. Who had this error?

Arek Wawrzyniak

File "E:\___SECourses\Ultimate_Video_Processing_v7\STAR\secourses_app.py", line 11, in import cv2, gradio as gr, numpy as np, torch, torchvision ModuleNotFoundError: No module named 'cv2'

SECourses

install error. delete venv, run installer and email me cmd logs please : monstermmorpg@gmail.com - also huge VRAM and general update coming to the app soon hopefully

SECourses

hopefully the app will be fully updated and it will use way lesser vram. i am waiting main repo to be updated they said withing a week

Hoàng Giang Sơn Trương

RTX 5090 but out of memory :( Star & SeedVR2 both the same. Please guide me on the best way to setting.

David

Hi, in the next version can you add the ability to "Batch Upscaling" for the images too? Now we can do only with videos, thank you 🔝

SECourses

sure next version will have so amazing features hopefully. currently you can batch upscale. in the batch upscaling select Direct Image Upscaling checkbox

David

Ok ok thanks, I will try.

Darwin Giordano

I would be interested to know what you think about "magnific.ai" and, in your opinion, how do they achieve such a good image that they promote using other methods? thank you

SECourses

Well they have money to train their own model. But SUPIR and SeedVR2 are best public ones atm. SUPIR is still really good

Tyler

Hello! Is this still the best opensource video upscale option? I'm used to just upscaling in comfyUI and interpolating... Is this process better?

JP LONB

Hi Furkan, any chance we get an update on this app soon. I keep having a black output but it seems to be a known issue. Any v8 version soon to be expected? Does the v7 work for anyone locally (I have a RTX4080, VRAM 16GB)? Thanks in advance!

JP LONB

Thanks for your quick answer! Looking forward to the upgrade!

SECourses

you are welcome sorry for delay

Pewburrito

Hi Furkan; whats the minimum number of Vram I will need to upscale a 1080p video into 4k?

SECourses

Hi. soon it will be under 12 GB i think. I am waiting them to merge updates into main repo. maybe even lower

Pewburrito

With STAR too, or just Seed? I have an Nvidia 3060 12gb, and both error out or out of memory on anything above 240 or 360p. Even 360p is a risk. I don't dare try 720p. I really want to believe you are right, friend :) Maybe I am doing something wrong?

SECourses

it will come with SeedVR2. Almost there

Diggy Dre

Can I use this to upscale standard definition stuff into 1080p with my 4080?

SECourses

with deterministic upscalers. for others i have to upgrade the app which i plan to work on this weekend hopefully

daniel gontz

Hello, what is the path that STAR is expecting to find SeedVR2 models in? The downloader py file is downloading the models for SeedVR2 in SeedVR2/models. But Star seems not to know where to look for download_star.py: snapshot_download( repo_id="MonsterMMORPG/SeedVR2_SECourses", local_dir=f"SeedVR2/models" ) Log: Hugging Face cache folders set to: E:\ai\Ultimate_Video_Processing\STAR\models Warning: SeedVR2 modules not available: No module named 'src.models' Using STAR repository base_path: E:\ai\Ultimate_Video_Processing\STAR Skipping import of cpp extensions due to incompatible torch version 2.7.0+cu128 for torchao version 0.14.1 Please see https://github.com/pytorch/ao/issues/2919 for more info W1213 11:30:20.065000 85164 Lib\site-packages\torch\distributed\elastic\multiprocessing\redirects.py:29] NOTE: Redirects are currently not supported in Windows or MacOs. 11:30:21 - INFO - Diagnostic: Explicitly set StreamHandler level to INFO. 11:30:21 - INFO - Logger 'video_to_video' configured with level: INFO. Handlers: [ (INFO)>] 11:30:21 - INFO - Found last used preset: 'SeedVR2_Video'. Attempting to load. 11:30:21 - INFO - Successfully loaded initial preset 'SeedVR2_Video'. 11:30:21 - INFO - Setting seedvr2.model from preset: 'seedvr2_ema_7b_fp16.safetensors' 11:30:21 - INFO - Initial preset for dropdown: 'SeedVR2_Video' (actually loaded) 11:30:21 - INFO - Scanning for image upscaler models in: E:\ai\Ultimate_Video_Processing\STAR\upscale_models 11:30:21 - INFO - Found 13 image upscaler model(s): ['2x-AnimeSharpV4_Fast_RCAN_PU.safetensors', '2xLiveActionV1_SPAN_490000.pth', '2xNomosUni_span_multijpg_ldl.pth', '2x_AniScale2_Omni_i16_40K.pth', '4x-AnimeSharp.safetensors', '4x-UltraSharp.safetensors', '4x-UltraSharpV2.safetensors', '4xNomos2_hq_dat2.safetensors', '4xRealWebPhoto_v4_dat2.safetensors', 'HAT-L_SRx4_ImageNet-pretrain.safetensors', 'Kim2091-4x-UltraSharp.safetensors', 'RealESRGAN_x4plus.safetensors', 'RealESRGAN_x4plus_anime_6B.safetensors'] 11:30:21 - INFO - [INIT] Starting SeedVR2 model scan... 11:30:21 - WARNING - [INIT] No SeedVR2 models found during scan 11:30:21 - INFO - [INIT] Converting SeedVR2 filename 'seedvr2_ema_7b_fp16.safetensors' to display name 11:30:21 - INFO - [INIT] Using first available choice as default: No SeedVR2 models found 11:30:21 - INFO - SeedVR2 Model Dropdown - Config model: seedvr2_ema_7b_fp16.safetensors, Choices: ['No SeedVR2 models found'], Selected: No SeedVR2 models found 11:30:22 - INFO - Gradio App Starting. Default output to: E:\ai\Ultimate_Video_Processing\STAR\outputs 11:30:22 - INFO - STAR Models expected at: E:\ai\Ultimate_Video_Processing\STAR\pretrained_weight\light_deg.pt, E:\ai\Ultimate_Video_Processing\STAR\pretrained_weight\heavy_deg.pt 11:30:22 - INFO - CogVLM2 Model expected at: E:\ai\Ultimate_Video_Processing\STAR\models\CogVLMFixed 11:30:22 - INFO - Initialized with default GPU: GPU 0: NVIDIA GeForce RTX 5090 (31.8GB) (GPU 0) 11:30:22 - INFO - Effective Gradio allowed_paths: ['C:\\', 'D:\\', 'E:\\', 'E:\\ai\\Ultimate_Video_Processing\\STAR', 'E:\\ai\\Ultimate_Video_Processing\\STAR\\logic', 'F:\\', 'G:\\'] * Running on local URL: http://127.0.0.1:7861 * To create a public link, set `share=True` in `launch()`. 11:30:23 - INFO - Initializing SeedVR2 tab, keeping existing model selection Please advise

SECourses

i am working on much better version of this. please give me few more days. star model is downloaded here snapshot_download( repo_id="MonsterMMORPG/Video_Upscale_Star", local_dir=f"STAR/pretrained_weight" )

daniel gontz

Thank you for the prompt response. My question was where the app is looking for SeedVR2 models. That I could not find. But yes, I'll wait for the new version. Thank you

SECourses

sorry for delay. new app will be hopefully 100x better

Michael

Can I run this install.sh in the same pod that is hosting my swarmUI? Will there be any conflict?

SECourses

Yes you can. Also this app will be massively updated hopefully soon. I am working it. A massive new app with much better performance and features. Sorry for delay