InfiniteTalk V3版 - 一键生成自然对口型数字人视频 图片说话/图片唱歌 支持50系显卡 ComfyUI+WebUI 一键整合包下载

song***

无言以对 发表于 2026-4-24 22:16
显卡太差的一次不要生成太长，重新运行生成短点的

我的显存16G，运存32G的，我只用了10秒的音频。

无言***

songtso8 发表于 2026-4-25 06:58
我的显存16G，运存32G的，我只用了10秒的音频。

复制控制台信息最后几行贴出来

scx***

失败了，报错


[15:33:54] [Gradio] * To create a public link, set `share=True` in `launch()`.
[15:34:56] [Gradio] [15:34:56] [Gradio] 开始生成视频
[15:34:56] [Gradio] [15:34:56] [Gradio] 图片已上传: 96v.jpg
[15:34:56] [Gradio] [15:34:56] [Gradio] 音频已上传: 录音.m4a
[15:34:56] [Gradio] [15:34:56] [Gradio] 视频帧数分配: 节点1=81, 节点2=0, 节点3=0
[15:34:56] [ComfyUI] got prompt
[15:34:56] [Gradio] [15:34:56] [Gradio] 任务已提交: da4f12d1-ef19-453e-b0af-bd15532dd2bf
[15:34:56] [Gradio] [15:34:56] [Gradio] 目标输出节点: 177 (帧数: 81)
[15:34:57] [ComfyUI] Using pytorch attention in VAE
[15:34:57] [ComfyUI] Using pytorch attention in VAE
[15:35:12] [ComfyUI] VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
[15:35:24] [ComfyUI] unexpected audio encoder: ['project_hid.bias', 'project_hid.weight', 'project_q.bias', 'project_q.weight', 'quantizer.codevectors', 'quantizer.weight_proj.bias', 'quantizer.weight_proj.weight']
[15:35:24] [ComfyUI] Requested to load Wav2Vec2Model
[15:35:25] [ComfyUI] loaded completely; 13654.80 MB usable, 189.00 MB loaded, full load: True
[15:35:42] [Launcher] Detected service crash
[15:35:42] [Launcher] Shutting down all services...
[15:35:42] [Launcher] All services stopped
[15:35:42] [Launcher] Shutting down all services...
[15:35:42] [Launcher] All services stopped
Press any key to continue . . .

scx***

[15:09:20] [Gradio] * To create a public link, set `share=True` in `launch()`.
[15:10:40] [Gradio] [15:10:40] [Gradio] 开始生成视频
[15:10:40] [Gradio] [15:10:40] [Gradio] 图片已上传: 96v.jpg
[15:10:40] [Gradio] [15:10:40] [Gradio] 音频已上传: 6月4日.mp3
[15:10:40] [Gradio] [15:10:40] [Gradio] 视频帧数分配: 节点1=81, 节点2=0, 节点3=0
[15:10:40] [ComfyUI] got prompt
[15:10:40] [Gradio] [15:10:40] [Gradio] 任务已提交: ce55a5c8-c7be-4302-9746-1bce6c61e77b
[15:10:40] [Gradio] [15:10:40] [Gradio] 目标输出节点: 177 (帧数: 81)
[15:10:41] [ComfyUI] Using pytorch attention in VAE
[15:10:41] [ComfyUI] Using pytorch attention in VAE
[15:10:56] [ComfyUI] VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
[15:11:08] [ComfyUI] unexpected audio encoder: ['project_hid.bias', 'project_hid.weight', 'project_q.bias', 'project_q.weight', 'quantizer.codevectors', 'quantizer.weight_proj.bias', 'quantizer.weight_proj.weight']
[15:11:08] [ComfyUI] Requested to load Wav2Vec2Model
[15:11:09] [ComfyUI] loaded completely; 13654.80 MB usable, 189.00 MB loaded, full load: True
[15:11:28] [Launcher] Detected service crash
[15:11:28] [Launcher] Shutting down all services...
[15:11:28] [Launcher] All services stopped
[15:11:28] [Launcher] Shutting down all services...
[15:11:28] [Launcher] All services stopped
Press any key to continue . . .