Model Environments

RVCBench integrates many third-party voice cloning and TTS models. These projects often pin incompatible versions of PyTorch, Transformers, ONNX Runtime, tokenizers, or model-specific helper packages, so one global Python environment is not expected to run every model.

During internal development, the model environment files were maintained under:

/tealab-data/rjin02/AudioWatermarkBench/envs

For the public release, the same environment specs are included in this repository under envs/. Use the environment that matches the model you are launching.

Creating an Environment

The environment YAML files reference the repository-level requirements.txt. Run the commands from the envs/ directory so relative requirement paths resolve correctly:

cd envs
conda env create -f qwen3-tts.yml
conda activate qwen3
cd ..

Then run the corresponding benchmark command, for example:

python run_vc.py --config-name ots_vc/clean/libritts/qwen3_tts_ots

To update an existing environment:

cd envs
conda env update -f qwen3-tts.yml --prune
cd ..

Model-to-Environment Map

Model family	Environment file	Conda environment name
General benchmark / fallback	envs/audiobench.yml	audiobench
BertVITS2 / SafeSpeech surrogate	envs/bertvits2.yml	bertvits2
CosyVoice	envs/cosyvoice.yml	cosyvoice
F5-TTS	envs/f5-tts.yml	audiobench
FishSpeech	envs/fishspeech.yml	fishspeech
Fish Audio S2	envs/fishspeech.yml	fishspeech
GLM-TTS	envs/glm-tts.yml	audiobench
GlowTTS	envs/glowtts.yml	glowtts
Higgs Audio	envs/higgs-audio.yml	higgs-audio
IndexTTS	envs/indextts.yml	indextts
Kimi Audio	envs/kimi-audio.yml	kimi-audio
MaskGCT	envs/maskgct.yml	maskgct
MGM-Omni	envs/mgm-omni.yml	mgm-omni
MOSS-TTSD	envs/moss.yml	moss
OpenVoice	envs/openvoice.yml	audiobench
OZSpeech	envs/ozspeech.yml	audiobench
PlayDiffusion	envs/playdiffusion.yml	playdiffusion
Qwen3-Omni	envs/qwen3-omni.yml	qwen3-omni
Qwen3-TTS	envs/qwen3-tts.yml	qwen3
Spark-TTS	envs/sparktts.yml	sparktts
StyleTTS2	envs/styletts2.yml	styletts2
VALL-E	envs/vall-e.yml	vall-e
VibeVoice	envs/vibevoice.yml	vibevoice
XTTS-v2	envs/xtts-v2.yml	audiobench
ZipVoice	envs/zipvoice.yml	audiobench

Reproducibility Options

The checked-in Conda YAML files are the lowest-friction public interface. They make model-specific dependencies visible and let users install only the environment they need.

For stronger reproducibility, generate platform-specific lock files from these YAML files with conda-lock, or publish prebuilt Docker/Apptainer images per model family. Containers are usually the most reliable option for CUDA-heavy third-party inference stacks, while Conda or Micromamba specs are easier for users who need to adapt paths, CUDA versions, or local checkpoint locations.