Real-time inference

[a-ghorbani]

Is SyncTalk suitable for real-time inference?
Are there any stats about latency anf performance?
Are there any benchmarks or optimization tips for real-time use?

[oguzhandoganoglu]

Good question !!

[jinqiupeter]

It definitely can. On my RTX 3090 I can have ~50 fps if everything is loaded to GPU. Here is my example of realtime chatting using SyncTalk:

Synctalk.realtime.chat.mp4

[oguzhandoganoglu]

How do you send a request after training the model? Can you share the code? Is it in the form of a test in the documentation?

[jinqiupeter]

Lipsync is also very good:

qs_cn_half.mp4

[oguzhandoganoglu]

Video doesn’t work. Alsı could you please answer my question if possible

[oguzhandoganoglu]

How do you send a request after training the model? Can you share the code? Is it in the form of a test in the documentation? @jinqiupeter