akwmw/gui-for-pdf-to-tts-using-speech-t5

Gui for pdf to tts using speech T5

Description:

the goal is this project is create ui for reading adamic paper using tts model orianly i want to Retrieval-based-Voice-Conversion-WebUI https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI to maketts but i have found a way to decode the model so i just use mircosoft t5 model instend but i stiall have trouble using speecht5 to find out what dataset code for what voice and the speecht5 have limit 512 token so i chunk the speech to 512 part each origaly i wan to add all the function below but the project already take longer then i like so i cut most of the fuction that i dot know how to do that may take me over a week to figure out how to do also i just find out you cant make unti test (pytest) on gui i can tran to use auto clicker and some machine learning to do the auto testing but this will take another week so here is the rush finally result

nougat-ocr only work with gpu if you using cpu use pdpdf2

nougat-ocr only work with gpu if you using cpu use pdpdf2

and nouget give a cleaner text https://github.com/facebookresearch/nougat

speech t5 from mircosoft

https://huggingface.co/microsoft/speecht5_tts

• gui
• prograss bar
• regex
• option to use nougt

whishfull goal

• clean ui
• fast gui load time
• use Hypothesis for unti test
• use rvc model for tts to make human like voice
• find a way to load model outside of cache libary
• ask to intall hugging face model on first time set up
• add differnt voice option from cmu-arctic-xvectors