❗ Common issues ❗

❌ Error messages

Error	Solution
TypeError: 'type' object is not subscriptable	You're on python < 3.10, Use Python 3.10
TypeError: unsupported operand type(s) for : 'types.GenericAlias' and 'type'	You're on python < 3.10, Use Python 3.10
I cannot install TTS, no matching version found	You're on python > 3.10, Use Python 3.10
Please install tensorboardX: pip install tensorboardX	You can ignore this warning as it is not a requirement.

🔍 Missing models

Issue	Solution
Where does the HuBERT model go?	All HuBERT models download automatically to data/models/hubert
RVC models don't get downloaded.	RVC models have to manually be downloaded and installed, unzip them into a folder inside of data/models/rvc

❗ Other issues

Issue	Solution
Launching takes a long time, it's always installing packages	Launch with the --skip-install flag.
I want to host the app over my local network	Launch with the -l/--listen flag.
I want to use a different port	Launch with the --port xxxx flag where xxxx is the port you want to use.
I'm running out of VRAM (CUDA Out Of Memory error)	Launch with --bark-cpu-offload for much lower vram usage at the cost of a small speed decrease, use --bark-low-vram to reduce vram usage and increase speed at the cost of quality. And use --bark-use-cpu to fully run the model on the CPU, with no VRAM usage, likely at the cost of a lot of speed.
Models aren't loading. The loading stays frozen.	Something's wrong with the file path of audio-webui (Too long?) Launch with the --no-data-cache