Started training my own LTX 2.3 LoRA specifically for... character replacement, using it here in combination with D(AI)vid's LoRA to strengthen the effect while avoiding ghosting Nowhere near perfect yet, but given that it's generated in 155s with no controlnets it's not badshow more

I am really super happy to show you my... research that transform 3D volumetric capture of man capture with Kartel.ai and with env gaussian splatting generated with World Labs in a three.js player . And all that is "upscale" using Hunyuan vid2vid model with a Lora train on videos on the main actor. Hope the future of movie is here :) . Still need some work, but you can get the ideashow more

New open source LoRA Multi-Angles I created with fal... for Flux.2 🔥 Flux2.-Lora-Multi-Angles: give a camera angle in degrees → get your image from that view. To build this, I created a system that captures 72 positions from Gaussian Splatting to generate the training dataset fully automatically in a webgl viewer. ▶️ Test: 📥 Weights: Should I open source my WebGL viewer so you can train your own from gaussian splatting ?show more

One of my favorite details in DialKit is the... slider component. Smart snapping on click, responsive drag handle, ability to type to enter value too... it even has rubberbanding. It's not perfect yet, but it's the best I've seen for this type of control.show more

Yeaah with Kartel.ai and Jesse Wellens ! So training... Lora for character on Flux and also for Wan 2.1 14B ( using dataset video ) make a Huge difference. So here you are seeing a volumetric capture of Jesse Wellens that we did in at the spatial studio of Kartel.ai We use a a World Labs gaussian splatting to create a simple environment that we put with the volumetric capture in our own webgl viewer ( out soon for everyone ). After that, We use after this output in a ComfyUI workflow with Wan Fun control ( a bit similar Vace but its working for Wan 14B) with the double loras , the first to generate the first frame with flux, and the second one to guide the generation of Wan 2.1 Control. So we keep very good consistency of character and position, and creating amazing worlds :) ! Hope you will find it cool !show more

While I agree that Angry Birds: Slingshot Stories is... not a perfect replacement for Angry Birds Toons, I do like it for what it gives to the Angry Birds canon, like how in-universe the Angry Birds theme we all know and love it's a war/military song that the pigs have PTSD for.show more

This tool is literally Higgsfield AI but FREE for... good. It's called Wan2GP. A full AI video studio built specifically for people without expensive hardware. Runs on as little as 6GB of VRAM, even old RTX 10-series cards and 8GB laptops. Everything stays on your machine, no uploads, no caps, no watermarks. What you get in one app: • Text-to-video and image-to-video generation • The best open models built in: Wan 2.2, LTX-2, Hunyuan Video, Flux • A full browser interface with a queue system • LoRA support to customize any model • Mask editor and prompt enhancer included A 5-second clip generates in minutes on a mid-range gaming rig. No subscription, ever. 100% Free. Open Source.show more

We’re excited to introduce Text-to-LoRA: a Hypernetwork that generates... task-specific LLM adapters (LoRAs) based on a text description of the task. Catch our presentation at #ICML2025! Paper: Code: Biological systems are capable of rapid adaptation, given limited sensory cues. For example, our human visual system can quickly adapt and tune its light sensitivity to our surroundings. While modern LLMs exhibit a wide variety of capabilities and knowledge, they remain rigid when adding task-specific capabilities. Traditionally, customizing these models requires gathering large datasets and performing often expensive, time-consuming fine-tuning for specific applications. To bypass these limitations, Text-to-LoRA (T2L) meta-learns a “hypernetwork” that takes in a text description of a desired task, as a prompt, and generates a task-specific LoRA that performs well on the task. In our experiments, we show that T2L can encode hundreds of existing LoRA adapters. While the compression is lossy, T2L maintains the performance of task-specifically tuned LoRA adapters. We also show that T2L can even generalize to unseen tasks given a natural language description of the tasks. Importantly, Text-to-LoRA is parameter-efficient. It generates LoRAs in a single, inexpensive step, based solely on a simple text description of the task. This approach is a step towards dramatically lowering the technical and computational barriers, allowing non-technical users to specialize foundation models using plain language, rather than needing deep technical expertise or large compute resources.show more

i didnt expect to get here so fast with... Fable WTF IT'S JUST BUILDING THE FULL GAME AND ASSETS OUT IN FRONT OF ME (i am trying to create a dynamic world sim, D&D style, with my AI characters) in order to create variance and randomness while generating synthetic data to grow the world model, i'm building out an actual game and resource mechanic, reward (earning resources/food/gold), punishments (death, gold loss etc), economy on how these Boo characters interact with each other given their situations and feelings and then training LLMs per character based on all the data they've genned during this experienceshow more

Soooooo I've been working on a soft outline/glow shader... I'll be releasing soon. It gets applied directly to the model's materials, and is everyday usable performance wise! It's not 100% perfect, but pretty close :D This is in a scene with no bloom:show more

More coming soon ! I manage to integrate for... the first time volumetric 3D video made with Kartel.ai inside the AI world generated in gaussian splatting by World Labs !! I coded all in three.js! And it's possible to integrate elements on the fly ;) And so you can imagine soon what we will produce with that en AI : consistency of character and environment, relighting etc...show more

i genuinely don't know how nobody has made this... yet. "Make Memecoins Great Again" has quietly become the movement. it's literally the slogan for the whole comeback. it started with Ansem, now everyone is running with it. the edits are everywhere on TikTok, retail is slowly coming back, and for the first time in a while people actually want to make memecoins fun again. 200k views, 26k likes: it's bigger than one coin now, perfect runner for the whole movement.show more

You can now track file upload progress when using... Vercel Blob in all modern browsers and JavaScript environments. When I started working on this, I thought, "It's 2024; it's gonna be simple LFG!!!" I was so naive 🤣 There's no way to track upload progress with fetch. Instead, you need to use a combination of Readable/Transform streams along with an old friend, XMLHttpRequest, to solve that. Here's how to do it…show more

we've shipped an official MCP Server 🚢 I was... a bit skeptical of MCP at the start, but as soon as started using it i changed my mind. It's a lot of fun using tools like Cursor/Claude to build with supabase [details in thread]show more

Reza Pahlavi: If yesterday's reformers today decide that it's... no use, we'd better join with the people and have the same demands, they're welcome to join us. It's as simple as that. As long as, again, they're not criminally implicated in the murder of their own citizens, because that's a different story, and they will have to account for it.show more

Ok in just a few hours, I have converted... my Claude Skill into an entire Mac App that can be used by anyone.. > Sign in with Google (mail + cal) > Give it disk access > Voila! Right now it's using my own API keys, but will transition to BYOK. The future is truly wild💃show more

This has been circulating in expedition forums for weeks.... No location. No names. No paper. Someone with a camera walked up to an opening in the rock and started filming. Look at what's inside. It's not a bear. Bears have fur, and this doesn't. It's not a rock. Rocks don't have that shape. Whatever it is, it fills the entire cavity. Nothing behind it. Nothing beside it. Just this. The clip ends before anyone gets closer. Nobody does. What do you think is inside that cave?show more

Bruh I felt so embarrassed when the guy told... the other guy to not come in front of the camera. I specifically kept the camera in a corner so that no one could see it and get awkward. You could see the embarrassment on my face, although it was in a light mood as those guys keep pulling each other's legs at the Gym all the time. But I still made it clear so no one else feels that way because he was so damn loud with “oye camera hai”. You don't own the gym, it's a public gym, so chill.show more