How long did the model take to train?

jasonppy / VoiceCraft

Zero-Shot Speech Editing and Text-to-Speech in the Wild

Other

7.66k stars 749 forks source link

Open platform-kit opened 7 months ago

platform-kit commented 7 months ago

And on what hardware?

Just wondering what the cost of duplicating this might be.

jasonppy commented 7 months ago

Please find details in the paper. We use 4 A40 GPUs and the biggest model took a little over 2 weeks to train

boringtaskai commented 7 months ago

How if I use dual RTX3090 ?