ayanami_v4_500st is a Japanese voice model that was created to perform songs.

For her training, most of the lines from the original series were used, as well as a couple of songs performed by Ayanami's Japanese VA.

As you might guess, it took me 4 attempts to take into account all my previous mistakes and create the best voice model with the Japanese voice of Ayanami (at the moment).

The voice model was used mainly in changing the vocals of the songs, although it also copes with the usual dialogues acceptable.

also here are examples of the voice of the model:<audio controls src="https://cdn-uploads.huggingface.co/production/uploads/64b0da8976e34f778f8dda97/CwSKM8O_TgaHrDxO0Tr0I.mpga"></audio>

<audio controls src="https://cdn-uploads.huggingface.co/production/uploads/64b0da8976e34f778f8dda97/ZNvdI8vhjOs9Oomz78SF2.mpga"></audio>

<audio controls src="https://cdn-uploads.huggingface.co/production/uploads/64b0da8976e34f778f8dda97/9o4nqgiFBxH_ZGTTdbpni.mpga"></audio>

upd 01.08 Also I created and added new model Ayanami's voice from rebuilds. Now I made it cleaner and removed the crackle in the background

<audio controls src="https://cdn-uploads.huggingface.co/production/uploads/64b0da8976e34f778f8dda97/Dbysy1ubzGzZ6erVwX5Yq.mpga"></audio>

<audio controls src="https://cdn-uploads.huggingface.co/production/uploads/64b0da8976e34f778f8dda97/w4JuJ-qDXyK5t4hl1btsb.mpga"></audio>

upd 02.11 created a new model (hopefully the last one) based on dialogue from the original series. This time I created a model that could reproduce extremely high notes, which seemed quite difficult given Ayanami's voice. (Ayanami_Rei_NGE_fs)

<audio controls src="https://cdn-uploads.huggingface.co/production/uploads/64b0da8976e34f778f8dda97/KnFWC5Pvc4PgpBTXtMxHn.mpga"></audio> <audio controls src="https://cdn-uploads.huggingface.co/production/uploads/64b0da8976e34f778f8dda97/EUgC4ftGcstWmHiIk2MgH.mpga"></audio>