[Bug]: "Nothing here" message when folder isn't empty

[BlafKing]

Is there an existing issue for this?

• [x] I have searched the existing issues and checked the recent builds/commits

What happened?

When SDXL 1.0 is selected, it doesn't list any models incompatible with SDXL 1.0 So my Textual Inversions models for example, do not get listed when a SDXL 1.0 checkpoint is being used.

The current description is quite confusing, since it states that nothing is in the folder, instead of explaining to the user that no compatible models for SDXL 1.0 are found.

I think it would be a lot clearer if the info displayed correctly states that the folder isn't empty, but nothing shows up because nothing is compatible with SDXL 1.0, for me this was really confusing and I didn't realize that the folder was showing empty because of the selected checkpoint's base model version:

with SDXL 1.0 Base selected:

Steps to reproduce the problem

1. Select a checkpoint with base model SDXL 1.0
2. Have no compatible models with SDXL 1.0 in a directory
3. Folder empty message pops up while folder isn't empty

What should have happened?

I'm not exactly sure what the best solution would be, Maybe a check could be done to see if the folders contain any models and change the message according to the presence of models in the folder?

For example if models are found but Automatic1111 has no available models to select a message like this could be shown:

Otherwise maybe the easiest solution is to change the message all together to something more general like:

Anything that would indicate to the user that the current Base Model of the Checkpoint is causing the model selection window to be empty, not because of an empty folder.

Another solution could be to dim any models that aren't compatible to indicate to the user that the models are found but not compatible with the current checkpoint Base Model.

Sysinfo

sysinfo-2023-09-21-17-56.txt

What browsers do you use to access the UI ?

Brave

Console logs

Log is irrelevant since no errors occur but here is the log regardless:

Python 3.10.6 (tags/v3.10.6:9c7b4bd, Aug  1 2022, 21:53:49) [MSC v.1932 64 bit (AMD64)]
Version: v1.6.0
Commit hash: 5ef669de080814067961f28357256e8fe27544f4
Pulled changes for repository in 'C:\Program Files\Automatic1111 SD\webui\extensions\a1111-sd-webui-tagcomplete':
Already up to date.

Pulled changes for repository in 'C:\Program Files\Automatic1111 SD\webui\extensions\sd-civitai-browser-plus':
Already up to date.

Pulled changes for repository in 'C:\Program Files\Automatic1111 SD\webui\extensions\sd-webui-3d-open-pose-editor':
Already up to date.

Pulled changes for repository in 'C:\Program Files\Automatic1111 SD\webui\extensions\stable-diffusion-webui-images-browser':
Already up to date.

Launching Web UI with arguments: --xformers --port 6969 --listen --update-all-extensions --no-half-vae --enable-insecure-extension-access --medvram-sdxl --no-hashing
Tag Autocomplete: Could not locate model-keyword extension, Lora trigger word completion will be limited to those added through the extra networks menu.
Aria2 RPC started
Loading weights [None] from C:\Program Files\Automatic1111 SD\webui\models\Stable-diffusion\kotosmix_v10.safetensors
Creating model from config: C:\Program Files\Automatic1111 SD\webui\configs\v1-inference.yaml
Running on local URL:  http://0.0.0.0:6969
Applying attention optimization: xformers... done.
Model loaded in 5.2s (load weights from disk: 0.7s, create model: 0.8s, apply weights to model: 1.6s, apply dtype to VAE: 1.0s, calculate empty prompt: 0.9s).

To create a public link, set `share=True` in `launch()`.
Startup time: 18.9s (prepare environment: 4.1s, import torch: 2.9s, import gradio: 0.8s, setup paths: 0.8s, initialize shared: 0.2s, other imports: 0.6s, setup codeformer: 0.1s, load scripts: 3.8s, create ui: 1.2s, gradio launch: 4.3s).
Reusing loaded model kotosmix_v10.safetensors to load sd_xl_base_1.0.safetensors [31e35c80fc]
Loading weights [31e35c80fc] from C:\Program Files\Automatic1111 SD\webui\models\Stable-diffusion\sd_xl_base_1.0.safetensors
Creating model from config: C:\Program Files\Automatic1111 SD\webui\repositories\generative-models\configs\inference\sd_xl_base.yaml
Applying attention optimization: xformers... done.
Model loaded in 25.9s (create model: 1.9s, apply weights to model: 21.9s, apply half(): 0.1s, load textual inversion embeddings: 0.3s, calculate empty prompt: 1.4s).

Additional information

A similair bug report (https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/13101) exists but I wouldn't consider this issue fixed yet since the current "Nothing here" message is very confusing.

[OliverTempel]

I can confirm this. SDXL Model => only safetensor files will be listed in embeddings SD1.5 Model => only .pt files will be listed.

[RansomSpawn]

its the same for me except for that no embeddings are displayed if i use a xl model, doesnt matter what xl model. if i use a non xl model the embeddings are displayed included the xl embeddings. doesnt matter if its a .pt or a .safetensors file.

[paboum]

Same for me. Are you guys using Civitai Helper too?

[BlafKing]

Same for me. Are you guys using Civitai Helper too?

I'm not, it's not caused by an extension, this is just the way that SD-WebUI currently works, it only shows models compatible with the selected stable-diffusion model

[paboum]

Then I guess the relevant code is here: https://github.com/AUTOMATIC1111/stable-diffusion-webui/commit/6f0abbb71a3f29d6df63fed82d5d5e196ca0d4de