Adding in new Avatars with Ex-Human API Key

[SILKWEBAGENCY]

Hi, I have recently purchased API access to Ex-Human and created a talking head avatar I would like to use in this project although having a bit of difficulty plugging in the code they are supplying. This is a Next.js app with React so not sure what the issue is? I am adding in my new addition to the project code just as these others are using Ex-human although I have an API Key now in addition to the token. Everything else I have, I have generated the .mov and .png image files for the avatar. Is there anyone that can assist me with getting this up and running? I have all other components of this, just need to add in my Ex-human talking head. much thanks!

[JavaFXpert]

Hi April,

Without modifying the app to add your talking head, do the other talking heads work with your EX-HUMAN token?

On Fri, Nov 11, 2022, at 3:12 PM, April Smallwood wrote:

Screen-Shot-2022-11-11-at-2 16 50-PM https://user-images.githubusercontent.com/95254406/201422906-7eb3afec-bd55-46af-9282-0c81883b5b39.jpg Hi, I have recently purchased API access to Ex-Human and created a talking head avatar I would like to use in this project although having a bit of difficulty plugging in the code they are supplying. This is a Next.js app with React so not sure what the issue is? I am adding in my new addition to the project code just as these others are using Ex-human although I have an API Key now in addition to the token. Everything else I have, I have generated the .mov and .png image files for the avatar. Is there anyone that can assist me with getting this up and running? I have all other components of this, just need to add in my Ex-human talking head. bernard-finishedV1 https://user-images.githubusercontent.com/95254406/201423408-9de8843c-b8a9-4c8e-8e69-06d73cf0ba0d.jpg much thanks!

— Reply to this email directly, view it on GitHub https://github.com/JavaFXpert/talk-with-gpt3/issues/12, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACLZ5WWGHIMFC7NCE3TVIODWH2SBTANCNFSM6AAAAAAR544LYA. You are receiving this because you are subscribed to this thread.Message ID: @.***>

[SILKWEBAGENCY]

Hi James,

Yes they are working well, I am attaching a screen. Was trying to use the code for Mary to add in Bernard but also adding him into the group for male voices. I also sent a request to tech support for Ex-human hoping maybe they can deliver the code needed for this.

Best, April Smallwood

[JavaFXpert]

Can you try reverting changes and replace Mary with Bernard everywhere? Regards,James (Jim) WeaverOn Nov 11, 2022, at 9:12 PM, April Smallwood @.***> wrote:

Hi James, Yes they are working well, I am attaching a screen. Was trying to use the code for Mary to add in Bernard but also adding him into the group for male voices. I also sent a request to tech support for Ex-human hoping maybe they can deliver the code needed for this. Best, April Smallwood

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: @.***>

[SILKWEBAGENCY]

Sure, I will give that a try and then upload it to a Github repo to share.

[JavaFXpert]

Cool. Please remember to not commit any of your keys.Regards,James (Jim) WeaverOn Nov 11, 2022, at 9:26 PM, April Smallwood @.***> wrote: Sure, I will give that a try and then upload it to a Github repo to share.

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: @.***>

[SILKWEBAGENCY]

It worked! Whoo hoo! Thank you. I think I can take the customization of Bernard from here. Brilliant project, so excited to work with this interface. Appreciate it!

[JavaFXpert]

Awesome. Nice work, April!Regards,James (Jim) WeaverOn Nov 11, 2022, at 9:51 PM, April Smallwood @.***> wrote: It worked! Whoo hoo! Thank you. I think I can take the customization of Bernard from here. Brilliant project, so excited to work with this interface. Appreciate it!

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: @.***>

[JavaFXpert]

April,By the way, the ex-human folks tell me that it’s best to give the talking head relatively short messages, as long ones tend to produce an error, after which the head freezes. Clicking the app’s microphone button recovers though. I’ve noticed that there also occasionally seems to be an error condition even with some short messages. I don’t know whether it’s a certain ascii character or what, but if you encounter it, please help me diagnose that annoying issue:-)Regards,James (Jim) WeaverOn Nov 11, 2022, at 9:57 PM, James Weaver @.> wrote: Awesome. Nice work, April!Regards,James (Jim) WeaverOn Nov 11, 2022, at 9:51 PM, April Smallwood @.> wrote: It worked! Whoo hoo! Thank you. I think I can take the customization of Bernard from here. Brilliant project, so excited to work with this interface. Appreciate it!

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: @.***>

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: @.***>

[SILKWEBAGENCY]

Jim, I will be working with this and will try to do a QA and experiment on the freezing instances and how that may relate to ascii characters or the microphone recovery. Thanks!

[JavaFXpert]

Cool. Thanks April!Regards,James (Jim) WeaverOn Nov 12, 2022, at 9:17 AM, April Smallwood @.***> wrote: Jim, I will be working with this and will try to do a QA and expiriment on the freezing instances and how that may relate to ascii characters or the microphone recovery. Thanks!

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: @.***>

[abuelgasimsaadeldin]

Hi @JavaFXpert @SILKWEBAGENCY,

I've recently purchased the API access to Ex-Human, I have generated the .mov and .png image files for the avatar and for testing purpose I have renamed them as Mary.mov and Mary.png respectively and replaced them with the current Mary files.

The Avatar is now showing the Custom one, however, when it starts talking, the old animated Mary is brought back.

Any help in solving this issue with the lip syncing to use the custom avatar would be highly appreciated.

Thanks and regards, Abu

[abuelgasimsaadeldin]

@JavaFXpert @SILKWEBAGENCY,

The talking image seems to be cached somewhere, I've cleared my chrome chache, deleted .Next, .idea, package-lock.json files but all this did not seem to solve the problem. Any help would be highly appreciated.

[JavaFXpert]

When you ask it to say a new thing, does Mary or the new avatar say it? Also, was this working for you earlier?  (Silk Web Agency, correct?)Regards,James (Jim) WeaverOn Jan 11, 2023, at 3:08 AM, Abu @.***> wrote: @JavaFXpert @SILKWEBAGENCY, The talking image seems to be cached somewhere, I've cleared my chrome chache, deleted .Next, .idea, package-lock.json files but all this did not seem to solve the problem. Any help would be highly appreciated.

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: @.***>

[abuelgasimsaadeldin]

Hi @JavaFXpert,

I ask it new things and it's able to respond normally but using the original Mary Avatar instead of using my own Avatar (however the video without talking is showing my own Avatar).

Regarding your second question, no it was never working for me even earlier and has always been this way.

FYI: I've deleted the "Mary.png" file completely and it's still showing her face while talking, which is quite weird..

Regards, Abu

[JavaFXpert]

When you had Ex-Human create the talking head, did they give you API endpoint parameters for "bot_name" "voice_name"? Does the new talking head work outside of the Chat w/GPT-3 application? Perhaps put some print statements around the following code to make sure that the correct parameters are being passed in: https://github.com/JavaFXpert/talk-with-gpt3/blob/b4896faef6d2a856b4f1e1fffbe2425857d74850/pages/index.js#L1362

Have you added a voice ID for your new talking head, as in the following file, perhaps using Masahiro as an example, or are you going with the approach of renaming an existing avatar? https://github.com/JavaFXpert/talk-with-gpt3/blob/5f6d2eda9f8cd74e062c28c54219e9696a02a113/pages/api/voice_options.js

I know that the code isn't straightforward, as there are exceptions for naming voices, etc. I suspect that the problem isn't caching though, rather, that mary's parameters ("Mary" and "Mary-EN" if I recall correctly) are being supplied.

Regards, Jim

On Wed, Jan 11, 2023, at 5:47 AM, Abu wrote:

Hi @JavaFXpert https://github.com/JavaFXpert,

I ask it new things and it's able to respond normally but using the original Mary Avatar instead of using my own Avatar (however the video without talking is showing my own Avatar).

Regarding your second question, no it was never working for me even earlier and has always been this way.

FYI: I've deleted the "Mary.png" file completely and it's still showing her face while talking, which is quite weird..

Regards, Abu

— Reply to this email directly, view it on GitHub https://github.com/JavaFXpert/talk-with-gpt3/issues/12#issuecomment-1378559669, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACLZ5WTPRTCEM4JNOLXR2PTWR2FTDANCNFSM6AAAAAAR544LYA. You are receiving this because you were mentioned.Message ID: @.***>

[abuelgasimsaadeldin]

Hi @JavaFXpert,

Thanks for your reply, I've just checked out the API endpoint received from my Ex-Human (shown below) and it seems to be quite different from the one used in the index.js. The "exHumanEndpoint" is entirely different and there isn't any "voice_name" either (I might be missing documentation of the "generate_lipsync" endpoint).

I've just recently subscribed to the Ex-Human API access after knowing that it was used in this Chat w/GPT-3 application, I have yet to test it out separately outside of the Chat w/GPT-3, perhaps that might be a great idea to sort of understand it better.

Lastly, I've added a voice ID to a new Avatar which I've recently added and created new instances of the Avatar everywhere that Mary was located (Mostly in index.js and voice_options.js). The video shows the correct Avatar, however, when it comes to the TTS, it does not work and the Avatar disappears (even though .PNG file is present).

Hopefully I've provided sufficient information.

Regards, Abu

[JavaFXpert]

When I received the talking heads from Ex-Human, they also gave me some code that describes how to use them with the API for the endpoint https://api.exh.ai/animations/v1/generate_lipsync

If you didn't receive it, please contact @.***

On Wed, Jan 11, 2023, at 9:58 AM, Abu wrote:

Hi @JavaFXpert https://github.com/JavaFXpert,

Thanks for your reply, I've just checked out the API endpoint received from my Ex-Human (shown below) and it seems to be quite different from the one used in the index.js. The "exHumanEndpoint" is entirely different and there isn't any "voice_name" either (I might be missing documentation of the "generate_lipsync" endpoint).

image https://user-images.githubusercontent.com/68045710/211802392-2b3b84d8-41a6-466d-842d-d92b6ba261bb.png

I've just recently subscribed to the Ex-Human API access after knowing that it was used in this Chat w/GPT-3 application, I have yet to test it out separately outside of the Chat w/GPT-3, however, I feel that would be a great idea to sort of understand it better.

Regarding the print statements, I've added them to make sure that the correct parameters were passed in, however, upon running of the command npm start I am greeted with example prompts from "intent_matching" and am unable to view the console.log() results in the logs (apologies as I'm still new to JS)

Lastly, I've added a voice ID to a new Avatar which I've recently added and created new instances of the Avatar anywhere that Mary was present (Mostly in index.js and voice_options.js). The video shows the correct Avatar, however, when it comes to the TTS, it does not work and the Avatar disappears (even though .PNG file is present).

Hopefully I've provided sufficient information.

Regards, Abu

— Reply to this email directly, view it on GitHub https://github.com/JavaFXpert/talk-with-gpt3/issues/12#issuecomment-1378885960, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACLZ5WUZUGCDPXVU7N57SATWR3DBFANCNFSM6AAAAAAR544LYA. You are receiving this because you were mentioned.Message ID: @.***>

[abuelgasimsaadeldin]

Hi @JavaFXpert Thanks for the info, I've emailed the Ex-Human support already and am currently waiting on a response. In the meantime I've printed out the bot_name and voice_name that are currently being called when running my custom Avatar and they are showing Hiroto and Hiroto-EN respectively which explains the change in Avatars during speech..

I've created my own custom talking Avatar from the Ex-Human Platform with the name "Debra", but that did not seem to work when passing it to the API Endpoint (even after adding it to the voice_options.js and index.js files). The Avatar just disappears when it starts to speak (but it does show Debra and Debra-EN in the logs).

• My question is, are the "bot_name" and "voice_name" derived from the API itself?
• Why does it show the Hiroto Animated Avatar while speaking when I am passing my API key to access the Ex-Human API and I clearly do not have that Avatar in my Ex-Human workspace nor do I have the .MOV/.PNG files in my local directory? (they have been deleted already)

I would really appreciate your input.

Thanks and regards, Abu

[JavaFXpert]

My suggestion is to do some print statements at various points in the program to see when things are changing. This will also give you more knowledge of the program itself.Regards,James (Jim) WeaverOn Jan 12, 2023, at 4:47 AM, Abu @.***> wrote: Hi @JavaFXpert Thanks for the info, I've emailed the Ex-Human support already and am currently waiting on a response. In the meantime I've printed out the bot_name and voice_name that are currently being called when running my custom Avatar and they are showing Hiroto and Hiroto-EN respectively which explains the change in Avatars during speech.. I've created my own custom talking Avatar from the Ex-Human Platform with the name "Debra", but that did not seem to work when passing it to the API Endpoint (even after adding it to the voice_options.js and index.js files). The Avatar just disappears when it starts to speak (but it does show Debra and Debra-EN in the logs).

My question is, are the "bot_name" and "voice_name" derived from the API itself? Why does it show the Hiroto Animated Avatar while speaking when I am passing my API key to access the Ex-Human API and I clearly do not have that Avatar in my Ex-Human workspace nor do I have the .MOV/.PNG files in my local directory? (they have been deleted already)

I would really appreciate your input. Thanks and regards, Abu

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: @.***>