I used clean audio clips from episodes of the anime and used the F5 Zero TTS feature on Hugging Face to generate over 5 minutes of normal speech and trained it on Google colab for 250 epochs.

Tip: Lower the transpose number to get better accurate results, for a male to male voice between -6 to -12 is accurate.

Additional Details:

Tags: Game Freak, Nintendo, Anime, huggingface, ai-model

Related URL: https://huggingface.co/Ryanham1lton/Toxicroak/resolve/main/Toxicroak.zip?download=true

Download Link: https://huggingface.co/Ryanham1lton/Toxicroak/resolve/main/Toxicroak.zip?download=true Hosted on Hugging Face

Convert this model in EasyAIVoice

Voice accords

Main sonic traits

Selected-sample labels, grouped by what they describe.

Age

middle aged adult

elderly

young adult

Music use

adlibs

speaking

producer tag

Style/timbre

villain

deep

serious

Use case

game

social

podcasts

Sample deck

Lean and sweet spot

Pick a pitch, compare the Quality Score, then Run. Hold Ctrl or Cmd to select multiple pitches.

top 4.93 · avg 4.41

female

0

male

18

male base · pitch 0 Quality Score 4.86 · heard as male

Autorun

Run

male

-16 2.51 -12 3.86 -8 4.84 -4 4.93 0 4.86 4 4.85 8 4.47 12 4.57 16 4.69

female

-16 4.58 -12 4.57 -8 4.17 -4 4.69 0 4.69 4 4.77 8 4.49 12 4.45 16 3.44

heard female heard male Quality Score = estimated audio quality, higher is better Ctrl/Cmd-click = select multiple pitches

Similar voices

This voice model sounds like

Nearby voices from generated samples. Pick a pitch in the Sample Deck to refresh the matches.

Finding nearby voices for this pitch...