|
|
|
|
|
From: smrwilly
, 1 months, post #2041 |
Wan : 5 free videos per day -> 1 free video per day.
:(
|
|
From: guest (rex)
, 1 months, post #2042 |
Maybe I'm the only one with a thing for superhero transformations
but here you go.
https://hailuoai.video/share/ai-video/eX21nv7ngja8
|
|
From: guest (Zit)
, 1 months, post #2043 |
Framepack on Pinokio is better than anything else for long videos
|
|
From: guest (ron)
, 1 months, post #2044 |
is using pinokio for framepack better than using the one click
framepack installer?
|
|
From: guest (Yokai)
, 1 months, post #2045 |
@ron : The version available on Pinokio doesn't seem specific or
particularly optimized, so it should be the same thing.
|
|
From: guest (naumm)
, 1 months, post #2046 |
Local Startframe / endframe generation -> pika
https://pika.art/video/ff2f8c33-cde3-4c0e-b08b-18145ed082ea
https://pika.art/video/4e2eeb94-72e7-4767-90ce-4f6edc163f77
https://pika.art/video/64505196-6c2c-4fd6-b419-fdb86a4bf585
https://pika.art/video/0dfdea7f-9254-493a-88dd-c7061019858d
https://pika.art/video/6227c733-1260-4664-a219-79b834eaa9a0
NSFW
|
|
From: hskfmn
, 1 months, post #2047 |
Prepare for trouble...make it double-D!
https://hailuoai.video/share/ai-video/qvjApY8nLAzN
|
|
From: guest
, 1 months, post #2048 |
Some man disguise as Sofia Vergara
https://hailuoai.video/share/ai-video/V4JnkErzzyaz
https://hailuoai.video/share/ai-video/g8xYlxDlq3gO
|
|
From: bump13
, 1 months, post #2049 |
Kling can make gender transformation
https://v21-kling.klingai.com/bs2/upload-ylab-stunt-sgp/se/ai_portal_sgp_m2v_txt2video_1080p_std_v16/d16e0718-5567-40f7-87b9-61fc43b88bf0_video.mp4?x-kcdn-pid=112372
|
|
From: bump13
, 1 months, post #2050 |
just create a standard prompt
"bald man wearing short pants and he smiling. he sitting in the
chair at luxury apartment lounge room. as he stands up, he smoothly
transforms into woman have long hair, wearing tank top and short
pants and she smiling while talks."
|
|
From: guest (guest)
, 1 months, post #2051 |
Just found out this new generator that focuses on transformation,
called higgsfield. You can select special effect that you want, and
in our case it is morphskin.
https://higgsfield.ai/motion/6308a9a6-877d-4fb1-be21-481638833983/982ca9a3-ff1f-4002-aaf2-ea862a4235a8
|
|
From: guest
, 1 months, post #2052 |
Another one:
https://higgsfield.ai/motion/6308a9a6-877d-4fb1-be21-481638833983/67ac5ad3-bb33-4bf2-9dc4-0ecc5ca0ca1b
|
|
From: smrwilly
, 1 months, post #2053 |
Not bad, but only 2 free tries, it's short.
|
|
From: guest (clop)
, 1 months, post #2054 |
I want to make sure I understand: When you guys lament that there
are only a limited number of free videos generated per day, or it
keeps censoring, or it's hard to get the prompts right, are you
folks doing all this using their web site? Is no one running the
WAN AI on their own computer? I realize not everyone has the
computing power to run a latent diffusion model on their own
computer, but it's not that hard. You just need a graphics chip on
your computer, and I'm sure at least 20% of you have have that
(probably more like over 50%). Then you put WAN on your computer (I
like it better than Hunyuan, which is bulkier for the same benefit;
LTXV doesn't work as well). Then you can generate what you want.
The results will still be random at first, and you generate a whole
bunch hoping that one of them will be decent. But if you find
something like, you see what random seed was used to generate this,
and then put in the same random seed, adjusting your prompt as you
want, so you don't have to depend on the "luck of the draw". You
can put in LoRA's to adjust what you want, like "I want bigger
nipples" or "I prefer Asian faces" or something. There are LoRA's
you can download for free on huggingface.co and civitai.com etc.
It might be because you look at the full WAN AI and see that it
takes up huge amount of memory and disk space. But you can use a
quantized model (GGUF) which someone has already pre-optimized for
you. Like, I have wan2.1_i2v_480p_14B_fp16.safetensors sitting on
my computer taking up 30GB of disk space, and is way too big for my
laptop with 4GB of GPU RAM on my nVidia graphics chip; but I use
the wan2.1-t2v-14b-q3_k_m.gguf model (you only need a Q3 model; it
gives decent results) which is only 7.1GB on disk and runs on my
4GB chip. If your chip has less RAM you can use a model that has
been quantized down even further and it will take longer but at
least you get to control what's running. Mine take about 20 minutes
to generate a video, so I just queue up a batch of prompts
overnight, and in the morning there are a whole bunch of videos
waiting for me, which gets me up in the morning, in more than one
sense of the word.
Or maybe there's something else I'm missing, and there's some
benefit I haven't discovered yet to doing it over the web on
someone else's computer where they set all the rules.
If I get around to it I'll post some videos that I generated, which
aren't that great but I didn't have to generate dozens to pick on
the one I want. I guess one advantage of using a web site is that
you can share your videos on their website, but I didn't use a AI
website so I'll have to go figure out this MegaUpload thing and
post some videos.
|
|
From: guest (lobstomonos)
, 1 months, post #2055 |
I've been playing around with both WAN and Hunyuan locally. Got
some good results, some poor.Not tried it with a LoRA yet.
The big advantages are no content moderation (obviously) and the
ability to tweak aspects of your output (length, resolution,
framerate etc).
The main drawback is power consumption - my electricity bills have
increased by a not-insignificant amount. Another drawback is that
it takes longer to generate a video locally. Using a 4070TI, a 5
second 480P video takes around 20-30 minutes to generate with the
WAN 14B model depending on how many steps you want the generation
to take.
I'm just waiting for either of those models to introduce a
reference-to-video feature like other generators have, although at
that point I think my GPU might have a meltdown lol
|
|
From: guest (clop)
, 1 months, post #2056 |
Okay, got the upload done. So here's the link to a
home-AI-generated "hey, that nurse isn't quite what she seems to
be" sort of video.
https://mega.nz/file/A3cmBKrB#UJv3oW18sd_nXfqQHXE81X-9KleGEDwm2fDhuVT5eCs
But if that's too conservative for you, you can get her to show her
boobs:
https://mega.nz/file/tiNkUIwA#-xqNE4jV_HNYYCrPvTMle7brBQhuZjjKlhY-VC3Vq5U
But if that's still not risqué enough for you, hopefully two
completely naked women will do it for you:
https://mega.nz/file/sqt2BBIQ#sb6P1t7X0g5vdKF__RrknKn56NJzMqBzw8mLFOAdEpY
Here's the folder with the stuff, so you can poke around for other
samples I'll upload.
https://mega.nz/file/sqt2BBIQ#sb6P1t7X0g5vdKF__RrknKn56NJzMqBzw8mLFOAdEpY
(I moved some stuff around after I got the links above, but
hopefully the links should still point at the same files. Let me
know if the links arent working. If they're not, you might want to
just use the link to the entire folder, the last link above.)
Hopefully this inspires some of you to try more stuff and share
with us. I am very grateful for all those people who made this
thread have TWO THOUSAND entries over the past half year, and I
have OVER TWO THOUSAND FIVE HUNDRED little morphing clips from this
thread, so ... thank you! I'm sure if I started naming names, for
sure I'd leave someone out, but you know who you are! Go check out
ComfyUI and HuggingFace.co and CivitAI.com and go do what you do
best! Hope to see more great stuff.
|
|
From: guest (pakistan666)
, 1 months, post #2057 |
higgsfield is trash, is been 2 hours my video still in queue.
|
|
From: guest (Captain Pickle)
, 1 months, post #2058 |
@2045
One of the big problems with local video generators is that they're
not easy to install. Getting the models, and LORAs, and
dependencies going is not trivial.
Sure YOU might be able to do it, and it seems easy. But I and many
other people are morons. I've been trying to get WAN AI to work and
i keep getting truckloads of cryptic errors that i don't understand
and thus can't fix.
Wtf is a "missing sageattention" error? i don't know, and i cant
even begin to fix it. Is there some Sage guy i have say hello to
and bring them coffee? How the fuck do i get his attention?
I don't blame the devs for such complexity. I understand the space
is evolving so fast that its not possible to make a
photoshop-simple app to do it. So we have to rely on an
over-the-internet solution where someone else handles the ever
changing and evolving landscape and ensures their setup works. I'll
even pay them a few bucks for their expertise if its reasonable.
|
|
From: guest (Yokai)
, 1 months, post #2059 |
@2058
Well on Pinokio it's supposed to be pretty straight forward, if you
encounter an error or a bug maybe you should try to get some help
from the Pinokio community on Discord and submit your issue with
the logs file. Judging from what you wrote I think the problem
could be because of a bad install, so try deleting Pinokio
completely (there are some guides on the Github I think but just
uninstall it through Windows and even delete the folder located in
AppData, Roaming) and make a clean reinstall. Sage attention is
something that comes along the version of Wan in Pinokio, normally
you don't have to install anything manually and that's why there's
a script for the installation, so maybe the issue is with Windows
or a bad install. Just ask from help from the Pinokio community if
you can, I mean if it still refuses to run after a clean reinstall,
they'll ask you to post details about your issue so that they can
help you.
Alternatively some people also use SwarmUI to use Wan, and if
you're really adventurous there's still ComfyUI (which isn't very
"comfy" for beginners tbf). Regarding the LoRAs there are a certain
of AI website that host LoRAs for Wan, you should be able to find
those easily.
If you really can't run it locally, I think you can generate for
free (but limited) on some AI websites.
|
|
From: guest (Captain Pickle)
, 1 months, post #2060 |
Oops i meant to direct my comment at @2054, not @2045, I'm fucking
dyslexic today lol
but anyway, i'm trying pinokio, my errors were while using comfy
ui, its not easy to install.
|
|
|
|
|
|
Add a New Message to This Thread
Posting Guidelines
Primary
- ANY POST(s) MAY BE REMOVED at the discretion of moderators
for violation of this website's guidelines/rules or any law
applicable to this website.
- All discussions must be transformation related OR of interest to
the transformation community.
- Please try to keep posts appropriate for younger teenage eyes to
view (Things you'd be comfortable sharing with a 13 year-old).
- Do not 'attack' other posters. You may attack an individual's
comments but not the individual.
- Be polite, courteous, and respectful. Please keep the conversations
civil.
- While use of an online alias is ok. Do NOT pretend to be another
(or registered) user.
Intellectual Property
- Do not post direct links to complete copyrighted works. Links
provided by (or approved by) the copyright holder may
be exempt.
- Links to content hosted on websites that have a clear, and easy to
find, process for dealing with infringement(s); and a known
reputation for following through with said process, may
be allowed.
- Please read this site's views on copyrighted content on the FAQ
page.
- Do not plagiarize. Credit your sources.
Helpful
- For posts and material that are not safe for viewing in a workplace
or with younger eyes, please mark the thread or post as either, "MATURE CONTENT", "ADULT CONTENT",
or "NSFW".
- Please post clear SPOILER warnings if you intend to give away
details that would spoil the story for those who have not yet seen
the material, preferably by using the spoiler
tags (don't forget to close the spoiler text with the end tag),
like so: <spoiler>Text to be "hidden"
</spoiler>
-- For additional help, please see: HOW TO: Use Spoiler tags
message thread.
- When starting a New Message Thread, please include details about
your topic of discussion or request in the Subject
field to help draw in readers who are interested and avoid wasting
the time of those who are not. Also helpful in searching for the
topic/thread later.
- If you post a link, please provide some detail as to what it is.
Identify the material if possible.
- Please post in English, or include an English translation, when
possible.
- If seeking material / information, include the results of your own
efforts so that others can offer better help (and avoid wasting
time and effort on redundant searches).
- When starting a new thread to request material or information,
please begin the subject with "REQ:"
or "REQUEST".
Etiquette
- Do not post messages in ALL CAPITAL LETTERS.
- Do not attack a user for improper spelling and/or grammar. Not all
of our community uses English as their native language.
- Please do not stray (far) from the topic of the original post.
- If a poster provides their contact information for any reason, do
NOT clutter the Message Board by asking them to contact you. They
gave out their contact information for a reason.
- No begging. Ask nicely.
- Do not feed the trolls and spammers. Please ignore them.
- Do not keep asking for clips after others have already posted
information on where to obtain the material.
- Please have some patience. Not everyone visits the board daily.
Your post may not be seen (by someone with the correct information
or idea) for several days or weeks.
- No bumping. Bumping will be treated as intentional spamming.
Advertising
- Do not post advertisements.
- One exception is for products or services that relate to entries on
this site. And even then, only post an announcement once. You
should add links or content to the appropriate database entries. Do
not repeat this information to the message board. Doing so will be
considered spamming.
Miscellaneous
- This list is neither all-inclusive or all-exclusive in nature, and
is meant to guide everyone on appropriate content for and conduct
on this Message Board.
- The purpose of this Message Board is to share information about
transformation-related content, news, ideas, etc. Also acceptable,
are conversations that are of interest to the transformation
community.
- Moderation of posts for reasons not listed above are at the
discretion of the Trusted Users who moderate this website, for the
purpose of keeping things within the "spirit" of the owner's wishes
and the website's purpose.
- If you disagree with a moderation decision, please make an argument
as to why it should be reinstated. Use the posting guidelines to justify your point.
Do not attack the moderator.
[Edit this Page] | |
|
|
|
| |
|
|
|
|