fudan-generative-vision / hallo

Hallo: Hierarchical Audio-Driven Visual Synthesis for Portrait Image Animation
https://fudan-generative-vision.github.io/hallo/
MIT License
9.49k stars 1.3k forks source link

Mandarin checkpoint released (中文权重开源) #194

Open DBDXSS opened 2 months ago

DBDXSS commented 2 months ago

First, we would like to express our gratitude to the authors for open-sourcing this outstanding work, which has demonstrated exceptional performance in English video generation tasks. Building on this project, we constructed the Mandarin dataset JDH-Hallo and made several adjustments and fine-tunings to the model. Both the model and checkpoints are now open-sourced and available for access at https://jdh-algo.github.io/JoyHallo/.

首先,我们十分感谢作者开源这项杰出的工作,它在英语视频生成任务中表现出色。在此项目的基础上,我们构建了普通话数据集 JDH-Hallo ,并对模型进行了多次调整和微调。模型和权重现在都已开源,欢迎访问 https://jdh-algo.github.io/JoyHallo/

APeiZou commented 3 weeks ago

mark

liutaocode commented 10 hours ago

First, we would like to express our gratitude to the authors for open-sourcing this outstanding work, which has demonstrated exceptional performance in English video generation tasks. Building on this project, we constructed the Mandarin dataset JDH-Hallo and made several adjustments and fine-tunings to the model. Both the model and checkpoints are now open-sourced and available for access at https://jdh-algo.github.io/JoyHallo/.

首先,我们十分感谢作者开源这项杰出的工作,它在英语视频生成任务中表现出色。在此项目的基础上,我们构建了普通话数据集 JDH-Hallo ,并对模型进行了多次调整和微调。模型和权重现在都已开源,欢迎访问 https://jdh-algo.github.io/JoyHallo/。

你好,感谢开源,请问你们stage1和stage2分别训了多久呢