Releases: mafiosnik777/enhancr
Release - 0.9.9 + Free Version 🎉
enhancr Pro 0.9.9 is available for Silver and Gold Patrons on Patreon now
https://www.patreon.com/posts/enhancr-pro-0-9-84064916
Free version is available to download now as well.
Changes:
- Added fp16 i/o to RIFE
- Replaced deprecated cupy.cuda.compile_with_cache with cupy.RawModule
- Updated to Python 3.11, TensorRT 8.6.1, cudatoolkit 12.1, PyTorch 2.1.0 nightly & Vapoursynth R62
- Implemented SwinIR (TensorRT)
- Switched webm muxer back to ffmpeg
- Filtering out generic VapourSynth error now, due to people not getting that it's not an error
- Added RealESRGAN (DirectML)
- Removed blur on Windows 10
- Rewrote Queue Backend partially
- Implemented CAIN (DirectML)
- Added dynamic axes to RVPV2 -> extra step for exporting to onnx isn't necessary anymore
- Added DPIR (DirectML)
- Updated ThemeEngine to make UI a bit more consistent
- Implemented SCUNet (TensorRT)
- Small UI Changes overall
Bugfixes:
- Fixed Tiling for ShuffleCUGAN (NCNN)
- Changed optShape calculation for DPIR
- Fixed Subtitle Check when selecting mp4 output container
- Re-enabled CUDNN tactics in TensorRT (thanks nvidia)
- Fixed RIFE TensorRT engine build crashing
- Scaled down Preview encoder (This fixes most "[h264/h265/av1_nvenc @ xxxxxxxxxxxxxxx] No capable devices found" errors. See FAQ in wiki for more infos.)
- Fixed RVPV2 ONNX conversion
- Added padding to NCNN for resolutions non divisible by 8, due to color channels not being put together properly after padding inside model
- Fixed memory leak when reading media metadata that would occur when processing a lot of items in queue
- Fixed duration formatting in Media Info
- Discord Rich Presence properly resets to idle state when process is completed/canceled now
- Fixed custom pth ESRGAN models
- Removed fp16 i/o from RVPV1 (TRT) to make it work again
Release - 0.9.8 + Free Version 🎉
enhancr Pro 0.9.8 is available for Silver and Gold Patreons now
https://www.patreon.com/posts/release-enhancr-81864462
Free version is available to download now as well.
Changes:
- Preview gets encoded on hardware now instead of software (No more performance impact for live preview)
- Added realtime player with perfect support for audio, subtitles, fonts, attachments etc.
- Updated to TensorRT 8.6.0 (Starting with TensorRT 8.6, support for 2nd generation Kepler and Maxwell (900 Series and below) has been dropped. You will need at least a Pascal GPU (1000 series and up) and CUDA 12.0 + driver version >= 525.xx to run inference using TensorRT)
- Added upscale frameskip (Skips inference on duplicate frames, granting a major speed boost on animation)
- Switched muxer from FFmpeg to mkvmerge
- Added GMFSS Fortuna (PyTorch + TensorRT), the best Frame Interpolation Model for Animation available right now
- Updated FFmpeg to 6.0 from nightly, due to hardware AV1 being in mainline now
- Implemented fp16 i/o (in simple terms this means reduced bandwidth is used up, thanks to half floating-point RGBH input instead of RGBS)
- Removed waifu2x, because of subpar quality
- Improved UI and UX consistency when a process finishes
- Added ShuffleCUGAN (NCNN + TensorRT) (Over 2x speed* of ESRGAN Compact with CUGAN like quality at half the VRAM requirements that CUGAN would normally need, made by @styler00dollar)
- Removed old unneccessary audio and subtitle extractions -> Inference starts faster now
- Inference uses half of available CPU threads now, due to better performance
- While building an engine from onnx, trtexec gets properly killed now when cancelling the process
- Various small UI changes
*This comparison only applies to TensorRT, ShuffleCUGAN is around same speed as ESRGAN Compact in NCNN
Bugfixes:
- Fixed the key shortcut to export
- Dragging around text in user inputs and terminal doesn't lock up the app anymore
- Fixed "height not divisble by 2" when preview is enabled
- Switching from FFmpeg to mkvmerge should have fixed the error that muxing takes forever in some cases
- Fixed a rounding error for optShapes that would occur when converting to TensorRT Engine in Restoration Tab
- Progress bar text doesn't overflow anymore into other UI elements when input filename is too long & has Chinese/Japanese characters
- Fixed Thumbnail Caching issues in Queue
- Custom Model selection gets saved now and persists through app restarts like any other setting
- Rendering correct thumbnail when toggling queue tab now
- Fixed Scene Detection and Frameskip for GMFSS (TensorRT)
- Button for clearing the queue is hidden now, while a process is running
- Fixed UI breaking when pressing model tab on a specific page of settings
Contributors
Assets 3
Release - 0.9.7
Release 0.9.7 is available for Gold Patrons on Patreon now
https://www.patreon.com/posts/release-enhancr-79850513
Silver Patrons will get Access in 3 days.
Sorry for the long waiting time, I got hit really bad with covid and currently focusing on a new app/project for you guys.
Next update is gonna be huge, I promise.
Changes:
- Implemented RealCUGAN (TensorRT) (Disclaimer: You might need to use tiling for 1080p on 8GB cards, else you'll run out of memory)
- Implemented SwinIR (PyTorch) (It's really slow, but the quality is definitely worth it | Thanks to Bubblemint for providing the model btw)
- Updated waifu2x to use faster to use faster ncnn mlrt runtime (30% speed increase)
- Added Option to export as Frame Sequence
- Removed menubar that would appear when pressing "Alt" key
- Added window controls for project page on Windows 10
- Increasing UI scale by 0.10
- Added GIF Input support
Bugfixes:
- Fixed FFmpeg complaining about "height not divisble by 2" when sampled dimensions mismatch (This usually happens with DVD sources)
- WebM export should work as expected now, due to re-encoding to Opus when muxing
- Fixed Playing Videos/Opening Output Folder via Queue context menu
- Fixed Queue context menu glitching out when having other UI scale than default
- Added trailing ? to muxer to make errors less likely on final muxing step
- Fixed Subtitle & Input path check
Don't launch enhancr from the Installer by selecting the checkbox, haven't gotten around to fix the main app retaining Admin Privileges from the Setup yet. OAuth breaks with this.
If you did do it, just restart the app to fix.
Release - 0.9.6
Release 0.9.6 is available for Gold Patrons on Patreon now
https://www.patreon.com/posts/release-enhancr-78147760
Silver Patrons will get Access in 3 days.
This might be the most huge update yet, enjoy!
Changes:
- Implemented GMFSS Union TensorRT (Disclaimer: Engine takes a long time to build ~ 20mins for 1080p)
- Implemented CAIN RVPv2 model, trained by sudo (best CAIN model yet)
- Using Mica Material instead of acrylic blur on Windows 11 now
- Overhauled UI a bit
- Implemented UI rescaling (finally)
- Added DRM and Patreon OAuth
- Updated ffmpeg
- Set float32_matmul_precision to medium for small performance boost on GMFSS Torch inference
- Updated TensorRT to 8.5.2 (Unfortunately you have to reconvert all your engines, because engines serialized in older versions of trt aren't compatible)
- Implemented proper Threading
- Removed pre-release Popup on Startup
- Added Option to clear Queue
- Implemented Asar for UI (should feel a lot snappier now and less startup time)
- Added option to use custom 1x models for RealESRGAN in Restoration Tab
Bugfixes:
- Fixed RealESRGAN NCNN not working due to using onnxruntime instead of ncnn wrapper
- Scene Detection is working properly on GMFSS now (slight skill issue with placement of frameskip and scdetect)
- TensorRT inference should show up now systems with both integrated and dedicated Graphics
- Fixed Scene Change Sensitivity value not saving and persisting after software restart
- Timer in Discord Rich presence doesn't reset every 2 seconds while in inference step anymore
- Fixed wrong dimensions when trying to convert restoration models to engine that are not DPIR
- Fixed stuck progressbar and fps/eta counter when cancelling process manually
Release - 0.9.5
Release 0.9.5 is available for Gold Patrons on Patreon now
https://www.patreon.com/posts/release-enhancr-77207698
Silver Patrons will get Access in 3 days.
I hope everyone had a great start into the new year!
This update took quite a bit of time, because I needed to find a way to make CuPy and Cuda Toolkit working in a portable way, because they're needed for GMFSS.
Sorry for that! I can assure you the results GMFSS provides are absolutely worth the waiting time, though.
I only can say it was absolute pain, can't even count how often I recompiled Torch and fought with Debuggers and CUDA the last few weeks, only for it to work on my system but not on others when I sent it out to testers.
Anyways, it's ready now for you guys to enjoy!
Changes:
- Implemented true batch processing (you can drag multiple files or directories now into enhancr)
- Added GMFUpSS / GMFSS Union, the best Video Interpolation AI for animated content currently out there
- Added automatic conversion to ONNX for pth ESRGAN models (both are supported now)
- Injecting environment hook now and added own cudatoolkit with cupy and torch, making implementing almost any Video AI possible for the future
- Terminal doesn't get spamed with "Frame: xxx/xxx (x.xx fps)" anymore, neatly displays in one line
- Changed default Scene Change Sensitivity to 0.180
- Added "buildOnly" flag to TensorRT engine conversion, reducing engine build times a bit
- On request of someone a default project name was added, to prevent creating projects with blank names
Bugfixes:
- Fixed Settings not properly saving
- Correct enhancr version is being rendered in log now
- Fixed error when trying to mux files with data streams
- Project files on the welcome page only get rendered if they still exist now
- Fixed loading spinner not properly hiding after some errors
- Fixed optShapes error that would make onnx impossible to convert when disabling fp16
- Fixed Issue where new CAIN engines wouldn't convert because of a formatting error
By the way we have a wiki now, if you didn't see it yet: https://github.com/mafiosnik777/enhancr/wiki
Hotfix - 0.9.4
Hotfix 0.9.4 is available for Gold Patrons on Patreon now
https://www.patreon.com/posts/hotfix-enhancr-0-75911481
I won't count this towards early access because of the broken release yesterday, so 0.9.4 will be available to Silver Patrons in 2 days.
Changes:
- Added warning message if GPU driver doesn't support required NVENC API (Update your drivers to at least 522.25 or newer, if you want to use Hardware Encoding)
Bugfixes:
- Uncommented function that was necesseray to detect if queue has finished, effectively making UI unusable after running once (Queue works again)
- Project paths get cut off again if path is too long, preventing it from wrapping out of container
- Discord Rich Presence resets properly when stopping queue manually now
- Dimensions get stored in queue objects now -> No more weird issues with converting engine for wrong resolution with multiple queue items
- Fixed problems with integrated GPUs (Dedicated GPU gets detected properly in UI now)
- Fixed Issues with Hardware Encoding
Release - 0.9.3
enhancr 0.9.3 is out on Patreon now for Gold Patreons!
https://www.patreon.com/posts/release-enhancr-75886925
Silver Patrons will get Access in 3 days.
Changes:
- Removed TencentARC/AnimeSR Upscaler, due to poor speed & not very promising results
- Added RealESRGAN (NCNN) for Intel & AMD Support
- Added padding to resolutions that are non-divisible by 8
- Added window border on Windows 10
- Implemented Hardware Encoding for AV1 (RTX 4000 Series, AMD 7000, Intel Arc) and all other formats on other GPUs
- Unsupported engines on non-NVIDIA GPUs get hidden now by default
- Implemented Scene Change Sensitivity Setting
- Lot of under the hood changes, thanks to @IcedShake
Bugfixes:
- Fixed Patreon & Discord Buttons having broken/no links
- Preview works offline now (it pulled a dependency from network before)
- Removed verbose output from trtexec in Upscaling
- Fixed spcaces escaping path (sigh.. yet again)
- Discord Rich Presence resets properly now on completion
- Removed solid background on Welcome Page
- Fixed media header border-radius clipping off when using too long filename
- Queue item titles dont wrap into new line anymore effectively breaking ui
- Fixed inconsistent behaviour of mediainfo container
- Too long filenames dont clip out of progress bar anymore
Contributors
Assets 2
Release - 0.9.2
enhancr 0.9.2 is out on Patreon now!
https://www.patreon.com/posts/release-enhancr-75477003
Changes:
- Implemented TencentARC/AnimeSR Upscaler (will probably get removed in next update again, due to poor speed & not very promising results)
- Updated CAIN (TensorRT) - Stacking channels instead of width now, resulting in ca. 13% speed boost + no more need for PyTorch as dependency
- Overhauled Discord Rich Presence, displays live fps, percentage & engine now
- Streams don't get merged anymore when an error occurs + added proper error message
- Added Setting to change Cache location, where files are being temporarily stored while being processed
- Implemented Frameskip for Interpolation, which grants another performance improvement, especially on eastern Animation, due to containing a lot of static frames, which can be skipped by the AI now
Bugfixes:
- Fixed calling binaries from installation path with spaces (e.g user folders with sur- and last name)
- Recompiled CAIN (NCNN) filter, with proper paths -> CVPv6 (NCNN) works again
- Switched from ffms2 to lsmash, to mitigate some files not being able to be read
- Fixed installer not being able to add registry keys on updates
(This is the last update that will be released for Silver & Gold Patrons simultaneously, 3-Day Early Access for Gold Patrons takes effect now)