Drew Thomasson

What's an example command of using OCR in headless So I can make test for that in the workflow

@ROBERT-MCDOWELL Fantastic the OCR workflow test are working as seen by the outputs from the workflow that succeeded [audiobooks.zip](https://github.com/user-attachments/files/23486094/audiobooks.zip)

nice nice workflow is passing Only thing is when looking in the results for the output of the `XTTSv2FineTuneCustom` It's using the default xtts voice, not the BadCartmanSouthPark Voice [audiobooks.zip](https://github.com/user-attachments/files/23588885/audiobooks.zip)...

Yes #### Workflow successfully downloads custom model zip I confirm the custommodel zip contains the correct CartmanSouthpark `ref.wav` file @ROBERT-MCDOWELL Log proof of download in English XTTSv2 headless custom fine-tuned...

Once we push the next update I'll get around to fixing it Google company did another update

[colab_ebook2audiobook (1).ipynb](https://github.com/user-attachments/files/23610613/colab_ebook2audiobook.1.ipynb) Progress so far messing with v25 it seems It'll run on CPU but the GPU is not detected by E2A in colab 🤷

The error is just telling you that you're looking for the app.py in the wrong location Just look at the Dockerfile in the repo to see how the image was...

#### The Dockerfile that's used to construct the Docker image is in the repo ## Also The docker image `athomasson2/ebook2audiobookxtts` Is being deprecated and marked as legacy (as its naming...

I'm guessing the guide that everyone is using for unraid is just out of date or something?

Oh yeah tru, here's the Dockerfile [Dockerfile](https://github.com/DrewThomasson/ebook2audiobook/blob/main/Dockerfile) My bad 😅 Also if there is an out of date guide would anyone be willing to send a link to it? 😅