MohammedEhab20/vibe-voice-egyptian-cfg40

VibeVoice Egyptian Arabic — cfg_scale=4.0

Fine-tuned VibeVoice on Egyptian Arabic dialect (checkpoint-9160).

cfg_scale=4.0

Balanced quality. Best tested result — recommended starting point.

Stops cleanly at EOS. Slight loudness drift on scripts over 40 seconds.

Repo contents

File	Description
model.safetensors	Merged model weights (5.1 GB, single shard)
config.json	Model architecture config
tokenizer.json + friends	Qwen2.5 tokenizer files
preprocessor_config.json	Audio processor settings
voices/egyptian_male.wav	Reference voice for male speaker
voices/egyptian_female.wav	Reference voice for female speaker
samples/demo_cfg4.0.wav	Sample output at this cfg_scale

Backend usage

from huggingface_hub import hf_hub_download, snapshot_download
snapshot_download("MohammedEhab20/vibe-voice-egyptian-cfg40", local_dir="./model")
male_voice   = "./model/voices/egyptian_male.wav"
female_voice = "./model/voices/egyptian_female.wav"
# Run inference:
# python inference_from_file.py --model_path ./model --cfg_scale 4.0 ...