[Bug]: Train>PreProcess> Use Blip not working

[ashequrkhn]

Is there an existing issue for this?

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

What happened?

I tried captioning 10 images with the intention of training a LORA but nothing is happening. It nothing shows up in the command line, not even an error message.

Steps to reproduce the problem

1. Start up webui-user.bat
2. Train tab
3. PreProcess tab
4. Selected folder
5. check Use BLIP

What should have happened?

I was hoping it would run and generate the expected .txt files, but I got nothing. Waited ~ 30 mins thinking maybe it was slow.

Commit where the problem happens

commit: 22bcc7be

What platforms do you use to access the UI ?

Windows

What browsers do you use to access the UI ?

Google Chrome

Command Line Arguments

--listen

List of extensions

stable horde worker (not active at the time)

Console logs

venv "C:\Users\asheq\Auto1111\stable-diffusion-webui\venv\Scripts\Python.exe"
Python 3.10.0 (tags/v3.10.0:b494f59, Oct  4 2021, 19:00:18) [MSC v.1929 64 bit (AMD64)]
Commit hash: 22bcc7be428c94e9408f589966c2040187245d81
Installing requirements for Web UI

Launching Web UI with arguments: --listen
No module 'xformers'. Proceeding without it.
Loading weights [cc6cb27103] from C:\Users\asheq\Auto1111\stable-diffusion-webui\models\Stable-diffusion\model_1_5.ckpt
Creating model from config: C:\Users\asheq\Auto1111\stable-diffusion-webui\configs\v1-inference.yaml
LatentDiffusion: Running in eps-prediction mode
DiffusionWrapper has 859.52 M params.
Applying cross attention optimization (Doggettx).
Textual inversion embeddings loaded(0):
Model loaded in 15.6s (load weights from disk: 5.0s, create model: 0.6s, apply weights to model: 4.5s, apply half(): 2.0s, move model to device: 1.2s, load textual inversion embeddings: 2.2s).
Running on local URL:  http://0.0.0.0:7861

To create a public link, set `share=True` in `launch()`.
Calculating sha256 for C:\Users\asheq\Auto1111\stable-diffusion-webui\models\Stable-diffusion\povSkinTexture_v2.safetensors: 0eced149d042a6e2f403debb0943d208286005de3655e5113e18f5f010d8e6b1
Loading weights [0eced149d0] from C:\Users\asheq\Auto1111\stable-diffusion-webui\models\Stable-diffusion\povSkinTexture_v2.safetensors
Applying cross attention optimization (Doggettx).
Weights loaded in 24.2s (calculate hash: 16.4s, load weights from disk: 0.8s, apply weights to model: 5.8s, move model to device: 1.2s).

Additional information

No response

[waynoedm]

I'm having the same issue on two installs. It doesn't appear CLIP Interrogator is working. Verified with img2img "Interrogate CLIP", and in the Train pre-processor menu as "Use BLIP For Caption". I tried disabling all added extensions for shits and giggles to see if something was interfering.

Deepbooru functions as expected if that helps you.

[waynoedm]

Tried this morning on one of my machines, and it re-downloaded everything and appears to work fine now. I'll try on my other system later tonight.

[waynoedm]

Actually just tried. Working on both systems now.

[ashequrkhn]

Well give it another shot, maybe there was a bug on BLIP's end.

[linuxdevopscn]

Encountered the same problem, the image preprocessing did not work, the log did not report any errors

[SeBL4RD]

Same problem here. Can't use BLIP Captation. Infinite load