It looks softer than a regular lanczos resize. Looks less like an “AI” upscale and more like it could have been done with traditional filters, but is easier doing it with VEAI.Īs the description says, it doesn’t upscale, just anti-alias. It increases focus, but doesn’t tend to generate any artificial detail. It completely tries to denoise the entire frame, with absolutely no grain or “dither” like patterns. 8 seems to try to get things even sharper and in focus and also tries to denoise better than 7, but still leaves noise. 7 produces much sharper results but also can enhance noise. 6 has the weakest effect in terms of focus / denoise. Liking it a lot!Īrtemis HQ is similar to LQ in the sense that versions 6, 7 and 8 seem to be variations of the same model, while 9 is a completely new model. It generates a lot more fine “AI” detail, but also generates some “dither” like noise similar to Gaia CG v6. 7 removes a little more noise than 6 8 is slightly sharper in some areas and less noisy (at the cost of a little detail) than 7.Īrtemis LQ 9 appears to be a completely new model, with much improved results. Missing Apple Core ML files, affects 1x and 4x scale mode, missing one block size file for 1x and one for 4x, library should still functionĪrtemis LQ 6, 7, and 8 are mostly the same. Missing Open Neural Network Exchange Single Float (fp32) files, breaks 1x scale mode Missing OpenVINO files, breaks entire library Missing entire library files for Apple Core ML, breaks entire library Missing OpenVINO files, breaks entire libraries Models with missing / incomplete model files on the download server: Models visually indistinguishable from one another: Produces results visually indistinguishable from regular resize (spline, lanczos, etc.) Here’s my personal notes that I’ve taken regarding the published (i.e., live on the public server) models so far: Here’s a copy/paste of a post I made in the beta forum with some additional details: It’s probably not noticeable in final live video. In my testing, they are slightly sharper than the “beta” model files. According to developers, the “release” Artemis v9 model files have had an additional week of training. The “beta” Artemis v9 model files are no longer available as the “release” ones overwrote the “beta” ones on the download server, since they have the same file names. The following models files are missing / not available from the download server: Missing model files for download: These OpenVINO files are missing for the. VEAI will download the model files it needs automatically. Now Artemis-LQ v8 will appear in addition to Artemis-LQ v9 in VEAI. So for example, open artemis-lq-9.json, then change "version": "9", to "version": "8", and save the file as artemis-lq-8.json. BE CAREFUL not to overwrite your original JSON’s. Change the number in line 3 "version": "X", to the version you want, then save the JSON as a new file, nameing it to the corresponsing version. Your current JSON files should be located in C:\Users\Admin\AppData\Roaming\Topaz Labs LLC\Video Enhance AI\models. Then, open your current JSON file in a text editor (i.e., Notepad). To get these files, first, make sure VEAI is closed. Here are all of the current live models that may or may not work for you (try them and see) for VEAI 1.8.0: (Some users have asked Topaz to switch to Vulkan instead of DX12 which would work on Win7.) The new engine also requires DirectX 12, which means no more ability to run on Windows 7. I believe this is because 1.6.1 and earlier versions used TensorFlow, and 1.7.x is a new engine that uses Open Neural Network Exchange. The older models (1.6.1 etc) will not work in VEAI version 1.7.x and up. Devs : Could you make all versions of models avaiable for download ?Īnd maybe label them as v4.0, 4.1, 4.2 aso ?įor example artremis v9.0 and artremis v9.1 ( the beta and the released one )
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |