yatesdr

Are you using a venv? I couldn't get it to install on base system since their requirements.txt is all jacked up. I recommend removing the specific versioning in requirements.txt and...

It uses about 6gb of vram for me making some short sentences, but it can vary depending on the samples and text length you're using.

I haven’t formally monitored for that, but it doesn’t seem to use a lot of regular ram. It may spike when it’s loading the model to vram, but I’d be...

Sounds higher than my install is running, what OS are you on? I'm on Debian 11, no X servers or anything just a ssh prompt. But honestly it could get...

Go in the requirements.txt file and remove the specific versions for everything. It's all messed up when they froze the versions for whatever reason. Just had the same issue with...

The issues I had with wheel were solved on 3.9 and 3.10 by updating pip. python3.x -mpip install --upgrade pip I think there's something going on with the wheel package...

If you make a new voice, you have to run "python3 setup.py install" again after adding your samples. And make sure to put them in the right voices folder.... it...

I had this issue too, but was able to solve it by putting the clips in ./tortoise/voices/snoop and then running python3 setup.py install. It found the voice and let me...

You have to install lm-sys/Fastchat and rename train_freeform.py to train.py. That should get it working if you're otherwise capable. I managed to get a merge done today with the following:...

I managed to merge the 30B today by doing the following: using python-3.10 venv git clone im-sys/Fastchat pip install . in Fastchat directory rename train_freeform.py to train.py install other missing...