Description

Hello. I'm sharing my AI model of Mie from the anime 'The Girl I Like Forgot Her Glasses,' which is currently airing and has 10 episodes. The model is capable of singing and speaking. First, I downloaded all 10 episodes and cut out the parts where the protagonist speaks. The dataset consists of 18 minutes and 7 seconds of audio that went through UVR with Voc FT + De Reverb processing. Next, I divided the dataset into about 110 files using audiosplitter. I used the Harvest model for training, which took 1 hour and 28 minutes for local training, and I used RVC 2. The model has undergone 280 epochs and has been trained for 8.1k steps.