Home
Artists
Search Recent Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register
Partychan Telegram Live Sex ThePornDude
Home Artists Posts Import Register

PLUGIN UPDATE 🎄 Alive v64 (files) (Patreon)

Published:
2023-12-24 13:42:48
Imported:
Tags:

Content

Merry Christmas! Thank you for the donations and helping me build cool stuff in 2023!

MORE INFO
- Some of these features are at their first iteration in Alive and they might be a bit rough around the edges, I only have so much time I can put into this. Sex poses for example work best with Alive Characters, they require a male & a female characters to be in the scene, in the bottom bar to have the pink & blue indicators (dots) on their avatars. With vams there's lots of ways to change things, skins, presets, etc and there are going to be cases where Alive v64 fails to identify the male atom correctly, it will say 'male atom not found' even though there is one in the scene. With Alive Characters that shouldn't happen though. Also refreshing the plugin should make it would with generic vam models as well
- for the Aesthetics app you have to select a layer in order to apply effects on it (opacity, color, etc). You can add multiple layers (normal vam decal textures) and they will appear at the bottom in a list. There you can select them and modify them individually.
- There's a new setting to disable Alive automatic features except for face expressions. It's in Alive Settings > Automatic Behavior. Male models by default have that turned on. Quite a few people DMd me with problems related to that, about how to turn off male automatic behavior because it interferes with VAM posession & other plugins. People that want to keep automatic body movement enabled on male characters too, or want to disable it on female characters, can change that here:



CHANGELOG



CREDITS

  • klphgz - poses ( from Crouching, Standing, Lying & Prone, Sitting packs)
  • Ashauryn - morphs ( from AshAuryn.Expressions.2, AshAuryn.Pose_Tools.2 and AshAuryn.Sexpressions.2, AshAuryn.AshAuryn's_Female_Genitalia_Collection.2, AshAuryn.AshAuryn's_Assorted_Expression_Pack_with_75_Morphs.7, AshAuryn.AshAuryn's_Tears_and_Pain_Pack.3, AshAuryn.AshAuryn_Sexpressions_2_Point_0.4, AshAuryn.Expressions.5, AshAuryn.Pose_Tools.4)
  • Ascorad - morphs ascorad.asco_Expressions.12.var
  • Jackaroo - morphs ( from Jackaroo.JarModularExpressions.2)
  • kemenate - morphs (kemenate.Morphs.10)
  • geesp0t - spankings vr hand collision detection (Spankings Plugin Mod v3.6 on reddit)
  • vamx,VAMDeluxe - poses (vamX.VAMasutra.1)
  • VRAdultFun, ElkVR - BVH Player 1.1.3


INSTRUCTIONS



DEPENDENCIES

ALIVE should work without these vars too but they're recommended for the best experience. They're all free VAM addons.



FILES

SPQRAlive.v64.var

Files

Virt-A-Mate: Alive v64 preview

Trailer for the v64 update of ALIVE for Virt-a-Mate. More info at: https://www.patreon.com/spqr_aeternum VAM CREDITS skin: PicoVam, Riddler, Xspada Jean Shorts: Riddler Shadow Saber Outfit: Re-visions VAM Iris reflection: Oeshii lashes: VL_13 hair: Roac PostMagic VFX: MacGruber SONG: madbello - Lie More & Adore https://soundcloud.com/madbello/latin-soft Lie More & Adore By madbello is licensed under a Creative Commons License. AI Voice (API integrated into Alive): https://elevenlabs.io/

Comments

Alex

This is amazing! Thank you, and Happy Holidays!

Heinrich

You're the best, Merry Christmas!

friendly neighbor

Ohh my Christmas present has arrived! I am always glad to see a new Alive update. Tbh pretty much my only reason to still start VaM.

Berzerker

Love your work merry Christmas

Dakoto

Thanks mate for this huge Christmas gift :) Happy holidays !

HelloFox

Hey, is it normal that the response time from the AI is so long? Or how can i make it faster? When im trying it on the webui its relatively fast. But ingame its rly slow. Up to 10 sek per answer. Thank you.

SPQRAeternum
>>125420264

Hello! I didn't notice anything like that but I'll look into it. Technically it should be the same speed. It might be slower though due to vam taking up GPU memory. That is if you're comparing it to the webui running when vam is stopped. If that's the case, on desktop turning on in vam settings Vsync (to limit the framerate) should help a bit. Turning down vam settings in general, limiting the number of lights and their shadows, the hair physics, glutes physics, clothes physics should help too. Keeping down the number of models in the scene as well. If vam is taking up a lot of the GPU, the AI will be considerably slower. Also, in the webui, when loading the model, setting a smaller context value I think should improve speed. Most models used to be 2048 tokens context until recently, some now go to 32k or higher. This controls with how much text history the AI works with (1 token = 4 letters). As the chat log adds up the speed decreases

Anonymous

Hi. Can I use a service like runpod running ooba remotely with your plugin?

SPQRAeternum
>>125535461

Hello, As long as the text generation ui API is accessible to the pc running vam, it should work with Alive even if the AI is running on a network pc or online. Alive only connects to it through the API, so where the AI is hosted shouldn't matter. Somebody else was just asking me about runpod, here's how he managed to get it to work: "The solution was switching from RunPod's proxy (the default) to using the alternative 'Through TCP Public IP' detailed on this page: docs.runpod.io/docs/expose-ports From a browser tab, the {pod_ID}-{internal_port}.proxy.runpod.net URL functions exactly the same as the public IP alternative. However, it looks like only the latter plays nicely with Alive. The successful URL was constructed using the Public IP with the port added to the end (not the 5000, but the new port it gets routed through.)"

Brent Johnson

Amazing trailer there! very nice updates. That's hilarious the shoe tying conversation.

Anonymous

Anyone having trouble with the sex poses being yanky?

Anonymous
>>125697920

Basically if i select a pose it still needs some adjustments. Maybe that´s normal? idk Let´s say usually i need to move the female slightly into some directions for good alingment. and the penis tip being often too low and weirdly positioned

SPQRAeternum
>>125697920

If you use standard models, the poses should align. If the models have body scale changes or exaggerated morphs there will be misalignments, so that might be one thing to cause problems There's no automatic alignment in Alive yet for sex poses. There are free vam plugins that might help with that, like DiviningRod or BodyLanguage. I might add some stuff like that too to make vam easier to play, I'm not a fan of having to manually pose models neither

David

I've been having a lot of fun with this, but now my AI girl has started to reply with the three dots, or incoherent replies; any ideas what I can do to correct this?

SPQRAeternum
>>125708449

Hello, there's a filter that might cause that. If the model starts giving censored messages/lectures when chatting nsfw stuff, Alive will try to block that and override the answer with ..., otherwise if that message becomes part of the log it makes the AI stay stuck in that narrative and you never hear the end of it. It can be turned off in Alive Settings > Chat > Messages cleanup. That controls more cleanup though and you'll likely get worse responses in general with that off. The chats will deteriorate over time depending on the AI models. Clearing or trimming the chat should help. Current AI interaction is meant like an episodic/session thing, we're not yet at long-term conversations. The best model that I would recommend now is this one: https://huggingface.co/TheBloke/dolphin-2.6-mistral-7B-GGUF/tree/main with TheBloke/dolphin-2.6-mistral-7B-GGUF and dolphin-2.6-mistral-7b.Q4_0.gguf in the fields in Oobabooga to download it. When loading it, you have to put the n-gpu-layers field to something like 16 and try it. If the model answers too slowly, increase that and load again, like 24,32. And so on until you find a value that's right for you. That controls how much GPU the model should use. If it's on 0, it's entirely on CPU and it will answer slower. The n_ctx value should be low like 2048 or 4096, for low memory use. Higher values makes the AI keep track of longer texts (more replies), at the cost of memory. With those settings this AI model should be very fast and it should be also smart enough to keep conversations longer. This is for the medium performance class anyway. I don't try larger AI models myself. An alternative is also https://huggingface.co/TheBloke/openchat-3.5-1210-GGUF (openchat-3.5-1210.Q4_0.gguf). For low performance models, I'd recommend https://huggingface.co/afrideva/MiniChat-1.5-3B-GGUF/tree/main (minichat-1.5-3b.q4_k_m.gguf)

David
>>125708449

Thanks for the quick reply; I'll try these solutions.

David
>>125708449

That seems to have done the trick! I've got a 3090 GPU and have 128gb RAM so I've set the n-gpu @ 32 and the n-ctx @ 8192 to see what happens! So far so good - using the dolphin-2.6 LLM. Do you have plans for long-term conversations because I think that would really add to the immersion! Thanks for your hard work!

SPQRAeternum
>>125708449

Thank you. I have lots of ideas but we'll see. I kind of lost support in 2023 by working on AI for VAM. Not to mention time. So I'm not exactly enthusiastic about AI updates (for virtamate). I just add them here and there. I make them out of flow, not plans

Mr Me

Great piece of work SPQR. Thanks. If you speak german in the chat and use ö, ä, ü and ß, the model understands you well, but the answer always contains question marks such as drau?en instead of draußen or draussen. This causes pauses, hiccups and broken sentences. oobaboogas webui uses the characters correctly in the answer. Am I doing something wrong?

SPQRAeternum
>>126059128

Hello, somebody else reported the issue on youtube and I prepared a fix that should be in the next release

Mr Me

Great. I'm looking forward to the next version. Thanks!

Smumblepooch

This seems like a great tool in the arsenal: https://github.com/myshell-ai/OpenVoice

SPQRAeternum
>>126431410

It's noncommercial use only unfortunately. They say they'll allow commercial use soon. But the company seems to be a bit weird, they're trying to make a crypto coin https://docs.myshell.ai/tokenomics. They'll likely tie it in with that somehow

SPQRAeternum
>>126431410

"A variety of optional tasks are proposed each season ranging from basic to more involved contributions. By completing tasks such as interacting with bots, providing feedback from different areas, or sharing their experience to help others join, users earn seasonal points which can then be redeemed for Shell Coins token rewards." lol this is gonna be a bit controversial..

Smumblepooch
>>126431410

Ooh, yikes. We should start FapCoin! You earn them by.. um.. well, we'll work that part out.