Hello People,
Im pretty sure that all of you have seen Ai made videos in internet one way or other. I decided to take this as a break from my main project to relax myself. Even though I end up not relaxing much.
This was a unique experencing, there were times I was happy to be born in this era that I could just easily ask anything I want an answer and get it without having to check documentations and stackoverflow(Always is and will be the goat).
This took me a little more than a week with a lot of procastination. I was generously helped by a lot of Ai tools on the way. I told myself that I would not use AI to complete this readme file, so do bear the gramatical mistakes and my poor choice of words and life decisions.
Coming to the actual scructure.
we have the, model.py (to load the local LLM model)
prompt_template.py (Making the prompt template)
Script_generation.py (Text generation using llamaCpp-python & Langchain)
text_to_speech.py (Just as the name suggests)
subtitles_generation.py (No need of explanation)
video_assembly.py (This is basically putting everything together to work)
video_uploading.py (The automation for video uploading)
check.py (For good coding practices)
main.py (obviously)
The maximum difficulty I felt was easily making the llama-ccp-python run on GPU. Had to make custom cuda-wheel without even knowing what it means. Then comes the video assembly part, which was only completed using the help of AI. Almost same with the subtitles generation part. In short, there was a lot of learning and difficulty in this mini-project. I would suggest you all to try it on your own too. Would be such a fun experinece to be honest.
Lets come to the boring part now.