i've reading at huggingface and other place about finetuning the language model. unfortunately this is above my horizon of technical knowledge.

i wonder if it would be possible/desired to integrate an option of finetuning the model within storytoolkit itself, to make this easier for "beginners"?

at the same time, as far as i understand, finetuning doesn't result in a huge gain compared to the available large models. as i work with non-english languages and speakers of slang, I would still like to give it a try.

for example, entering a corrected phrase in the start box works perfectly to correct the word all along the interview. so in theory, making the model learn some expressions would help a lot in the transcription of numerous interviews.