VibeVoice_7B_hun_v2

VibeVoice_7B_Hun_v2

This is my newest finetuned VibeVoice 7B (Large) model tailored to Hungarian language. I trained LoRA for the LLM module, performed a full-finetune on the Diffusion head modules, and merged each of them into the base model.

To finetune the model I used the following code.

Thank you for JPGallegoar for that amazing VibeVoice trainer!

Training

This model was trained on RunPod cloud GPUs.

Inference

For inference, you can use

• this Comfyui node
• Demo codes on VibeVoice Community's repository

Examples

These examples were made with 4bit inference. One can get even better results without quantization.

Sample 1

"Az utcák lassan megteltek emberekkel, ahogy a város ébredezett.
A kávézók teraszain gőzölgő csészék mellett beszélgettek az emberek, miközben a villamos csilingelve gördült el a sarkon.
A levegőben friss péksütemény illata keveredett a tavaszi széllel.
Minden arra utalt, hogy egy nyugodt, szép nap veszi kezdetét."

<div style="display: flex; gap: 20px; align-items: center;"> <div style="flex: 1;"> <strong>Fine-tuned Model</strong> <audio controls src="https://huggingface.co/Cseti/VibeVoice_7B_hun_v2/resolve/main/assets/test_withmodel_1.flac" style="width: 100%;"></audio> </div> <div style="flex: 1;"> <strong>Base Model</strong> <audio controls src="https://huggingface.co/Cseti/VibeVoice_7B_hun_v2/resolve/main/assets/test_no_model_1.flac" style="width: 100%;"></audio> </div> </div>

---

Sample 2

"Őrült űző üldöző őz őrjöngő őrült őrzőjével ügető ürge üvöltözött.
Örvénylő örömök örökös özönével ölelő ösvényeken ődöngő űrlények ütköztek össze önérzetesen."

<div style="display: flex; gap: 20px; align-items: center;"> <div style="flex: 1;"> <strong>Fine-tuned Model</strong> <audio controls src="https://huggingface.co/Cseti/VibeVoice_7B_hun_v2/resolve/main/assets/test_withmodel_2.flac" style="width: 100%;"></audio> </div> <div style="flex: 1;"> <strong>Base Model</strong> <audio controls src="https://huggingface.co/Cseti/VibeVoice_7B_hun_v2/resolve/main/assets/test_no_model_2.flac" style="width: 100%;"></audio> </div> </div>

---

Sample 3

"Csapzott cserecsapat cserebogár csapongott cserepes cseresznyecsokrok csücskében,
s közben csipcsup csipogással csipkedte csípős csipkebokor csúcsát."

<div style="display: flex; gap: 20px; align-items: center;"> <div style="flex: 1;"> <strong>Fine-tuned Model</strong> <audio controls src="https://huggingface.co/Cseti/VibeVoice_7B_hun_v2/resolve/main/assets/test_withmodel_3.flac" style="width: 100%;"></audio> </div> <div style="flex: 1;"> <strong>Base Model</strong> <audio controls src="https://huggingface.co/Cseti/VibeVoice_7B_hun_v2/resolve/main/assets/test_no_model_3.flac" style="width: 100%;"></audio> </div> </div>

Important Notes: This model is created as part of a fan project for research purposes only and is not intended for commercial use. The dataset I used might contain material, which are protected by copyright. Users utilize the model at their own risk. Users are obligated to comply with copyright laws and applicable regulations. The model has been developed for research purposes, and it is not my intention to infringe on any copyright.

Support

Producing and sharing this kind of open-source work requires renting cloud GPUs, which gets expensive quickly. If you find it useful and would like me to keep contributing, your support is very much appreciated: