Develog - 3 This week I have mainly worked on creating Ai for the updating scoreboard with just a screen shot of the game scoreboard and up to this time I have added only for e-football and I am planning to add specific for some games and one for general. in this journey I tried mainly to create the AI myself so I researched how to do then I came to the idea of like it extract text from the image and I have used these three text excretion library called Easyocr, Pyteseract and Paddleocr but this I tried but here the main problem I faced is that they just extract the text but some time that will not able to I identify some small text and from that I cannot set who is the winner then I came to an idea to create an LLM using Pytorch with these neural network but when I started to do I faced a problem that my laptop is a office laptop so I cannot make such kind of LLM by my own because this lap cannot handle it. so I decided to use the google ai vision tool so now know from the beginning but I just want to create my own but due to this problems I use google ai vision it good by using a specific model I got 500 request per day(RPD) and now I have integrated that withe frontend like when they upload the scoreboard it will say analyzing the match after the Gemini response and save data in Json file in a format with some extra details like which match came the result came then only it will redirect to the main page and now I need to add the scoreboard, XP-points and some special extraction of score for some specific game