DrewThomasson
commented
1 month ago It might be that for my import your suppose to be importing the folder containing all the files, and for some reason I had to make the vocab file have a after it like "vocab.json" instead of "vocab.json"
Confirmed on my end I get the same error if I incorrectly rename the "vocab.json" file to "vocab.json_"
Error when I get when I incorrectly name the vocab.json file
No API token found for 🐸Coqui Studio voices - https://coqui.ai Visit 🔗https://app.coqui.ai/account to get one. Set it as an environment variable `export COQUI_STUDIO_TOKEN=FIX:
- If the custom model vocab file In the model dir is named "vocab.json" currently, then rename it "vocab.json_"
- OR if the customer model vocab file in the model dir is named "vocab.json_" currently, then rename it "vocab.json"
-If you imported it into VoxNovel then your imported custom model folder should be within the folder VoxNovel/tortoise/voices/Jay Snyder.M/model/{Here is where the custom model for that voice is imported into}
That should fix your issue. :)
Ignore this below lol this is extra just in case, Your issue is MOST LIKELY the thing above. :)
If not
Try using my docker image to fine tune a model locally (If you have a Nvidia graphics card): https://hub.docker.com/repository/docker/athomasson2/fine_tune_xtts/general
If you don't have nvida graphics card you could just my google colab image tho: https://colab.research.google.com/drive/1sqQqzupo2pdjgggkrbM60sU6sBFYo3su?usp=sharing
ScratchMode
I've Imported this Clone a Voice Sample, and this Model and then when trying to generate audio immediately stalled out with the following error.
Using XTTS V2.0.2 Model Fine Tuned with this bastard of a git project