Independent researcher
Joined December 2018
- Tweets 28
- Following 82
- Followers 628
- Likes 1
9 Photos and videos
Pinned Tweet
Jun 1
I spent $30k and 3 months RL post-training an anime video model.
This is only step 30 out of a planned 1000 step run.
All samples are local text-to-video with no reference image/audio. Since it's based on LTX-2.3, each output takes under a minute on a single GPU.
I'm 19 and a solo researcher. Most of the budget went into ablations, reward design, and trying different configurations before reaching this setup.
The run is still extremely early, but the results already look much better than I expected.
It's compute-limited, not idea-limited.
I'm starting a company to continue scaling this and build frontier stylized video models.
If you're an investor, compute partner, video team, or someone who wants to help build this, DMs are open.
36
34
378
23,340
Jun 8
I've got an idea to get Seedance 2-level conditioning flexibility on a limited budget. A training recipe that relies on synthetic data and a very recent post-train method that is neither RL or SFT.
My goal is to reach reference capability and quality parity much, much cheaper in training and inference cost.
Gonna build it. I'd really appreciate any help in compute or funding.
1
6
786
2 Mar 2025
Looking for a co-founder to build the next generation of waifu tech! Figured out the solution to create a new interactive experience but struggle with app dev.
Kind of the downside of focusing too much on ML.
Ideally someone with mobile/web and maybe some cloud ML experience.
5
1
14
1,465
24 Jul 2024
I've set out some conditions for any future solvers to add:
• No implicit solvers: those require root-finding, meaning a Jacobian has to be computed by running a backward pass through the model during LBFGS optimization.
• No non-RK methods: this would cut off linear multistep methods like Adams-Bashforth or Adams predictor-corrector. From my testing, RK methods perform better, even for explicit RK vs. predictor-corrector linear multistep.
• No duplicate methods: if two methods have different coefficients, they aren't duplicates (scipy methods have different coefficients and solver implementations).
That means the current 31 solvers are almost all that exist to satisfy the conditions above.
Project's done! I need to figure out what to make next.
2
997
24 Jul 2024
Last major update!
• Added solver settings for adaptive_scipy
• Adaptive solvers now show the number of steps taken
• Accurate 𝜎 timestep info is now displayed
Check out the most comprehensive fixed and adaptive higher-order samplers on ComfyUI!
github.com/wootwootwootwoot/…
3
22
2,610
22 Jul 2024
Refactored and fixed some bugs with the progress bar!
Also wrapped the solvers from scipy.integrate
If you count the a-methods as 2 (since they work with both the adaptive_pid and fixed_scheduled controllers), then this node has (excluding forward euler)
31 new samplers!
I also tried the implicit solvers and they didn't work.
Every implicit solver has a root find step, and that takes forever to converge.
That leaves 3 new methods from scipy: se_RK23, se_RK45, and se_DOP853.
I think this node has the most new working samplers for ComfyUI (a for adaptive, f for fixed, s for scipy, e for explicit).
7
692
21 Jul 2024
While trying to push the CFG scale up, I implemented some Explicit RK solvers for ComfyUI
- 10 new adaptive step samplers
- 8 unique fixed step samplers (excluding forward euler)
- Best new sampler (perhaps) -> fe_ralston3
Check it out!
github.com/wootwootwootwoot/…
2
10
85
6,512
21 Jul 2024
the new class of models idea didn't work out well, so i tried this instead (which works decently well)
5
601
17 Jul 2024
diffusion models are definitely still not dead. sure, optimal transport conditional flow matching is provably better, but so much of the community was already built on discrete time diffusion.
and with kolors out (an SDXL model trained with DDPM formulation and eps-pred objective). i doubt the switch from diffusion to OT-CFM will affect the quality as much as the other techniques shown in the technical report.
if they made kolors work with the SDXL architecture, then it's shown that hybrid transformer-UNets are still competitive. they might just not scale as well as pure DiTs.
1
17
1,407
17 Jul 2024
i have an idea for a slightly modified class of SDXL models that would mostly be compatible with existing finetunes and loras
it's been proven to work well on SD1.5 with good results
definitely next on my bucket list
will post updates and releases soon, hopefully, if it works
2
16
949
15 Jul 2024
i'd like to continue working on anime animation tech. version 1 is designed to be distilled for realtime inference. version 2 won't be concerned with realtime inference and would probably be based on a flow-matching mmdit with more fine-grained control and even better quality.
2
25
1,050
15 Jul 2024
training is done in 2 days!!!
as for inference compute requirements: it's basically the same burden as animatediff
will probably work on realtime inference next month after figuring out life. realtime txt/vid2vid on distilled models is already empirically shown to be possible.
1
9
992
14 Jul 2024
yo @EsotericCofe @anifusion_ai
wanna join forces
best manga tool best character animation model = ???
12
28
254
25,611
4 Jul 2024
my linkedin got restricted (for, using an anime pfp, presumably) so this is honestly great
4
11
1,658
3 Jul 2024
i've just realized it's been training for 40 days on 4xA100 SXM. it's going to take a bit more, around two weeks. after that, i'll try to get it to run in real-time (it's actually viable with the architecture).
i think the hardest part about this project is funding, not as in compute, but for personal stuff. i'm saying this as a broke college student who has access to a research cluster.
i'm curious. if i open a patreon, would i have enough to pay off rent and stuff and keep working on it and other future projects without financial worry? maybe when it's done i can dip my toes into becoming a vtuber in the process too.
you know, i've always wanted to create a startup that does research pertaining to anime. i wouldn't know how to get started now though.
the other option would be if any individual/startup/company is looking for someone to join their team and is willing to pay me.
dms are open for any inquiries or ideas.
meanwhile, here's another run with the same old checkpoint. keep in mind that parts of the architecture are still very undertrained and that results in inconsistencies and flickering.
i can't turn the unconditional guidance up at the current stage for the same reason (the blacked-out regions are caused by this). it's also very sensitive to the controlling pose, but these issues are easily solvable after the fact.
33
46
315
32,482