Help with fine tuned accent

[GUUser91]

I'm trying to train a model with a fine tuned british accent. I could only gather about 3 minutes and 20 seconds of audio for the dataset. Here's the input file.
https://vocaroo.com/1gY2BK8MTiTF
Here's the reference file.
https://vocaroo.com/1bTycJauDp9i
Output file from a pretrained model
https://vocaroo.com/1cdag3gy5958
Output file from a fine tuned model
https://vocaroo.com/1fvqjU7YBGrI

Edit: Nevermind I got the fine tuned model to work by following the instructions from this link
#131 (comment)

In the config_dit_mel_seed_uvit_whisper_small_wavenet.yml config file, I changed the in_channels info from 768 to 1280

Then I replace the whisper-small info with whisper-large-v3-turbo. I suggest using whisper-large-v3 if you're having trouble.

Here is a output files with the improved fine tuned model
https://vocaroo.com/17qPajlqbHnK