Update: Distilled v1.1 is live

We've pushed an LTX-2.3 update today. The Distilled model has been retrained (now v1.1) with improvements to audio quality and a slightly refined visual aesthetic. It's available on HuggingFace alongside the previous Distilled version.

Along with the new checkpoint, we've also retrained the distilled LoRA, updated all four ComfyUI example workflows, and refreshed the union control and motion tracking IC-LoRA checkpoints to work with the new base model (these replace the previous versions in place).

No major architecture changes, just refinement across the board. Files are live now. Would love to hear your impressions, especially on the audio side.

And stay tuned, more updates are on the way.

https://redd.it/1skds12
@rStableDiffusion

Free AI Voice Cloning with Qwen3 TTS — Google Colab Notebook (works on free tier, no GPU needed)

I've been using Qwen3 TTS for a couple of months now and figured I'd share a Colab notebook I put together for it. I know most of you have probably seen the model already, but setting it up locally can be a hassle if you don't have the right GPU, so this might save someone some time.

The notebook runs on the free Colab tier, no API keys or anything like that — just open and run.

Colab notebook: https://colab.research.google.com/drive/1JOebp3hwtw8BVeosUwtRj4kpP67sBx35
GitHub: https://github.com/QwenLM/Qwen3-TTS
For local install without terminal, Pinokio works well too: https://pinokio.computer

___________________

Also recorded a walkthrough if anyone needs it: https://www.youtube.com/watch?v=QmfiU8V5xq4

https://redd.it/1skeqk5
@rStableDiffusion

Turning Anime into Real and testing Klein9b vs Qwen Edit 2511 (Workflow Included)

https://redd.it/1skaqt5
@rStableDiffusion

AnimaYume - Anima finetune.

AnimaYume is a text-to-image model fine-tuned from [Anima](https://huggingface.co/circlestone-labs/Anima), a high-quality anime-style image generation model developed by [CircleStone Labs](https://huggingface.co/circlestone-labs). It builds upon [Cosmos 2](https://research.nvidia.com/labs/dir/cosmos-predict2/), a model developed by NVIDIA’s research team.

"For version 0.4:

* This version was trained on Anima Preview 3 using a custom dataset. In this release, I improved prompt understanding and artist style. Based on my testing, some artist styles match my expectations, although I haven’t tested everything in detail since I’m currently quite busy :<. Additionally, I fixed several issues from Anima Preview 3 that also appeared in Preview 2." [AnimaYume - v0.4 | Anima Checkpoint | Civitai](https://civitai.com/models/2385278/animayume?modelVersionId=2851312)





https://preview.redd.it/gf5sg4htezug1.png?width=2048&format=png&auto=webp&s=c749b214b11a6aefffedfe0c2751dfe4baa96953

[AnimaYume HF](https://huggingface.co/duongve/AnimaYume)

https://redd.it/1skfebq
@rStableDiffusion

LTX distilled lora 1.1 vs. 1.0 comparison

https://redd.it/1skjtz0
@rStableDiffusion

New WAN 2.2 Lightx2v speed lora 260412

Barely tested, hoping to get some feedback.

Official Full Model lightx2v/Wan2.2-Distill-Models at main

Scaled fp8 and extracted lora obsxrver/wan2.2-i2v-lightx2v-260412 at main

https://redd.it/1skkotf
@rStableDiffusion

I made a playable ping pong game where every frame is ai generated. This is my interactive diffusion model I made from scratch.

https://redd.it/1skmmnp
@rStableDiffusion

Does anyone recognize what artist tags the user @Magnus_waifu on Twitter/X might be using for their images?

https://redd.it/1skox6h
@rStableDiffusion

I found this interesting as it gives insight to how Z-image Turbo breaks down a prompt and then enhances it before image generation. Auto-translation to English included below in the text body.
https://huggingface.co/spaces/Tongyi-MAI/Z-Image-Turbo/blob/main/pe.py

https://redd.it/1skkd0a
@rStableDiffusion

LTX 2.3 how to stop Characters from "Cloning" themselves

https://redd.it/1skraz7
@rStableDiffusion

Work in Progress Encoder and Decoder!
https://redd.it/1sktnsk
@rStableDiffusion

Forget about VAEs? SenseNova's NEO-unify achieves 31.5 PSNR without an encoder – Native Image Gen is coming.

https://redd.it/1skv88l
@rStableDiffusion