[Bug]: multiple LoRa in prompt with PonyXL

DenSckriva commented 4 days ago

Checklist

[X] The issue exists after disabling all extensions
[X] The issue exists on a clean installation of webui
[ ] The issue is caused by an extension, but I believe it is caused by a bug in the webui
[X] The issue exists in the current version of the webui
[ ] The issue has not been reported before recently
[ ] The issue has been reported before but has not been fixed yet

What happened?

Hello !

I'm coming back to you for a problem that I can't resolve. I can only use one LoRa in my prompt.

If I put two, only the last LoRa in the prompt is used. I can no longer use several LoRa. I would like to point out that I currently use PonyXL models.

I of course tried to change the weight of each of them. But here again, only the modification made to the latest LoRa has any effect.

I have no problem with SD1.5. If you have a solution or ideas, I'm interested!

Steps to reproduce the problem

1 Type your positive prompt like : text, text, text, , text text, text, , text 2 Type your negative : text, text, text, (text:1.1), text 3 Generate

What should have happened?

Generation of an image with the two lora mentioned in the prompt

What browsers do you use to access the UI ?

Other

Sysinfo

sysinfo-2024-07-04-20-57.json

Console logs

Python 3.10.11 (tags/v3.10.11:7d4cc5a, Apr  5 2023, 00:38:17) [MSC v.1929 64 bit (AMD64)]
Version: v1.9.4
Commit hash: feee37d75f1b168768014e4634dcb156ee649c05
Launching Web UI with arguments: --medvram-sdxl --xformers --api --gradio-allowed-path G:\StabilityMatrix-win-x64\Data\Images
2024-07-04 23:09:33.756382: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
WARNING:tensorflow:From G:\StabilityMatrix-win-x64\Data\Packages\Stable Diffusion WebUI\venv\lib\site-packages\keras\src\losses.py:2976: The name tf.losses.sparse_softmax_cross_entropy is deprecated. Please use tf.compat.v1.losses.sparse_softmax_cross_entropy instead.

Loading weights [c762d00ce6] from G:\StabilityMatrix-win-x64\Data\Packages\Stable Diffusion WebUI\models\Stable-diffusion\zavyfantasiaxlPDXL_v10.safetensors
Creating model from config: G:\StabilityMatrix-win-x64\Data\Packages\Stable Diffusion WebUI\repositories\generative-models\configs\inference\sd_xl_base.yaml
Running on local URL:  http://127.0.0.1:7860

To create a public link, set `share=True` in `launch()`.
Startup time: 11.4s (prepare environment: 1.8s, import torch: 2.4s, import gradio: 0.6s, setup paths: 4.4s, initialize shared: 0.2s, other imports: 0.4s, load scripts: 0.5s, create ui: 0.5s, gradio launch: 0.2s, add APIs: 0.1s).
G:\StabilityMatrix-win-x64\Data\Packages\Stable Diffusion WebUI\venv\lib\site-packages\huggingface_hub\file_download.py:1132: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
  warnings.warn(
Applying attention optimization: xformers... done.
Model loaded in 6.0s (load weights from disk: 0.2s, create model: 1.6s, apply weights to model: 2.2s, apply fp8: 1.6s, calculate empty prompt: 0.3s).
100%|██████████| 40/40 [00:13<00:00,  2.93it/s]
Total progress: 100%|██████████| 40/40 [00:10<00:00,  3.75it/s]s]
100%|██████████| 40/40 [00:10<00:00,  3.82it/s]
Total progress: 100%|██████████| 40/40 [00:10<00:00,  3.69it/s]s]

Additional information

No response

nekoworkshop commented 3 days ago

Sounds like #15995

@DenSckriva Could you try again with FP8 disabled?

DenSckriva commented 2 days ago

It seems to work but I have to switch the VAE from "full" to "TAESD". Without FP8 it generates NaN errors and nothing else

AUTOMATIC1111 / stable-diffusion-webui