@_Envy_ It usually takes no more than 2-3 hours from the time of publication. However, it still doesn't work. There are several LYCORIS on the site that do not allow you to select them for generation. Especially those designed for 1.5 and XL at once.

This way don`t work online. As for zoom, there is his Envy Zoom Slider XL 01

@mihain Sry, i ment a viable zoom-out feature for landscapes, towns aso. Zoom Slider XL 01 works for portraits/people, but had its problems with "bigger" compositions (at least on my end)

@_Envy_ Hi Envy.

Your model classification for the CivitAI-website upload is correct (SDXL), yes. So it wasnt clear to me either, why the model didnt show up for SDXL accordingly.

Until, more or less by accident, I checked the MetaData baked into the downloaded LoRA-file itself in Automatic1111 a few minutes ago: aparently the classification of the base-model the LoRA was trained on (MetaData variable: "ss_base_model_version") is missing/wasnt written into the files MetaData by the training-app you used.

Automatic1111 just seems to default to SD 1.5 in this case, until you manually tell it otherwise.

In case of the CivitAI-generator and some other SD apps: if the software used is looking/filtering for the ss_base_model_version variable, the returned value would be "missing". At best, this would result in the LoRA not showing up in the options when using the "wrong" checkpoint. At worst, this could trigger a software-error while trying to load the LoRA - all depending on how lazy the writer of the SD-software was (speaking from coding-experience :D).

Here is the MetaData for your/this Model:

{ "ss_mixed_precision": "bf16", "sshs_model_hash": "786e1e80cb7a930b19f6eb124b6bdcf3ee5d28adf9ac4f21bec7c0902abde974", "sshs_legacy_hash": "308772ba", "ss_output_name": "EnvyPanoramaXL01" }

And, as an example, here is the MetaData from the xl_more_art-full-v1 LoRA:

{ "ss_base_model_version": "sdxl_base_v1-0", "modelspec.implementation": "https://github.com/Stability-AI/generative-models", "modelspec.architecture": "stable-diffusion-xl-v1-base/lora", "ss_network_module": "networks.lora", "modelspec.resolution": "1024x1024", "modelspec.sai_model_spec": "1.0.0", "modelspec.merged_from": "xl_more_art-full-beta2, realmix-lora", "ss_network_dim": "128", "modelspec.prediction_type": "epsilon", "ss_network_args": { "conv_dim": 32, "conv_alpha": 32 }, "sshs_legacy_hash": "c89f945d", "ss_v2": "False", "modelspec.title": "xl_more_art-full-beta3_1_0.5", "sshs_model_hash": "fe3b4816be8330f9902c28f6fe313ef42054c5cec0c2bb46480a5748cfbfd343", "ss_network_alpha": "128", "modelspec.date": "2023-08-31T02:14:18" }

The "ss_base_model_version": "sdxl_base_v1-0" information (or a variation of the version like 0.9) seems to be prominent in most SDXL LoRAs.

I dont know anything at all about training LoRAs, so the question would be: is there a checkbox/setting in the training-app you use, that would write the "ss_base_model_version" parameter and its value into the generated LoRA-MetaData information?

We are on the case (see the other comment) :D

Because this is LYCORIS not recognized by DT you can't install it using (Enter url...),
Download it locally (Download folder in Files) then install it using ( select from Files...) in DT LORA manager.