Text to Pose settings

[Tejaswini2612]

Hello,

I have been trying to reproduce results for text to pose, however, the settings provided on the git page are not working for text to pose.

Can you please share the settings and code( if it is different) for T2P and T2G2P?

Looking forward to hearing from you.

Thanks!

[divyachhipani]

Hi Tejaswini,

If you check the config file, you are able to set data>src as text or gloss.

Thanks,
Divya

[FangSen9000]

@Tejaswini2612 @divyachhipani

When I use the text setting, there is no video generation. When I use gloss, there is video generation. Do you have this situation?

[FangSen9000]

@Tejaswini2612 @divyachhipani

I have found that it seems that we also need to make a new glossary that the original code does not give. README doesn't specify that either

[hacker009-sudo]

@FangSen9000 please share the pretrained model, if you have stored it somewhere. Request you to please help it is really really urgent.

[khannabeela]

@FangSen9000

I am facing the same problem as @Tejaswini2612 and @divyachhipani,
Can you please tell me what do you mean by " make a new glossary that the original code does not give"? and would appreciate it if you give me any suggestions on how to solve this problem.