Perceptual LoRA Toolkit now supports Z-Image Turbo

https://redd.it/1ttak8l
@rStableDiffusion

lora dataset images and captions

Okay. I hear a lot of do and don't's, but *gawd* damn, I need more.

Character lora. 25 images. All 1024x1024, all consistent, varying, ...in my mind complete for at least a *functional* if not *flexible* lora.

How the hell do I caption this to be easy model side? I dont want to have to fine tune knobs and prompt engineer like Gemini and other llms are doing to my captions.

I have a highly toxic and inflexible lora iteration right now, I'm not dumb enough to require crash coursing, but im stuck.

I know the *"transient state"* of the image as a whole, including viewpoint should be tagged, but how does one ensure accuracy for the training of the character?

TriggerWord, camera angle, objects/lighting/background to not triggerword bake, but what ELSE about the character needs to be captioned for flexibility in the character themselves? I know clothing and accessories so a bunch of crap doesnt get welded to the character, but hairstyles and expressions?

Those *make* the character, but doesn't tagging them.... remove them from the character? .....but then dont all expressions and hairstyles get averaged and welded together?

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

Best anime model for multiple characters or Lora?

Hey everyone

I'm really struggling with multiple characters and their details. I can do a prompt that says 2girls but I sometimes get 3, etc. Or try specifying characters and their details and they'll be opposite Ori say I want small breasts and I get huge ones or something random. Or I get the perfect prompt and I regenerate it or read n later and it'll be f*cked. Is there something I can do/use in the prompt or a Lora I can use or something? I've tried pony, illumi, illustrious, noobai, wai-illustrious

Hope you can help

Regards

https://redd.it/1tt95u3
@rStableDiffusion

FLUX.2-klein-base-9B ControlLight LoRA Release for changing lighting of a photo
https://yfyang007.github.io/ControlLight/

https://redd.it/1ttfv5z
@rStableDiffusion

Nvidia releases Cosmos3-Super-Image2Video . 64B parametres
https://redd.it/1ttjuc0
@rStableDiffusion

Nvidia releasesCosmos3-Super-Text2Image model . 64 billion paramteres
https://redd.it/1ttjrip
@rStableDiffusion

The Cosmos omnimodel family of models - 3 variants Edge(4B) , Nano(16B) , Super (64B)

https://redd.it/1ttka77
@rStableDiffusion

An AI-generated short film I spent weeks creating.
https://www.youtube.com/watch?v=HCJ4ubbaCxU

https://redd.it/1ttk4as
@rStableDiffusion

Bernini released. Unified Video generation and editing model. Built on Wan-2.2

https://redd.it/1ttn2kd
@rStableDiffusion

Local AI News You Missed - May 2026

Releases you (might of) missed in May 2026:

**🧠 LLMs**

1. [**Supra-50M**](https://huggingface.co/SupraLabs/Supra-50M-Base) - A tiny model that packs a heavyweight punch in a small package.
2. [**MiMo-V2.5-coder-Q2**](https://huggingface.co/jedisct1/MiMo-V2.5-coder-Q2) - Supercharges coding and tool calls specifically for Macs.
3. [**Kezmark ErniePEUnleashed**](https://huggingface.co/Kezmark/ErniePEUnleashed) - A tool to help craft cinematic scene prompts.
4. [**OBLITERATUS Qwen3.6-27B-OBLITERATED**](https://huggingface.co/OBLITERATUS/Qwen3.6-27B-OBLITERATED) - A model fine-tuned to snip out refusal circuits completely.
5. [**Nemotron-Labs-Diffusion-14B**](https://huggingface.co/nvidia/Nemotron-Labs-Diffusion-14B) - Turbocharges text generation with three simple modes.
6. [**Tencent Hy-MT2-1.8B**](https://huggingface.co/tencent/Hy-MT2-1.8B) - A pocket-sized model for 33 language translations.
7. [**Tencent Hy-MT2-30B-A3B**](https://huggingface.co/tencent/Hy-MT2-30B-A3B) - A powerful 33-language translator that runs locally.
8. [**MiniCPM5-1B**](https://huggingface.co/openbmb/MiniCPM5-1B) - One model with dual modes for fast chat or deep thought.
9. [**G4-MeroMero-31B-uncensored-heretic**](https://huggingface.co/llmfan46/G4-MeroMero-31B-uncensored-heretic) - Slashes 85% of refusals for creators.
10. [**Gemma-4-Gembrain-31B-It-Uncensored-Heretic**](https://huggingface.co/llmfan46/Gemma-4-Gembrain-31B-it-uncensored-heretic) - Reduces AI refusals by 87%.
11. [**BitCPM4-CANN-8B**](https://huggingface.co/openbmb/BitCPM4-CANN-8B) - Slashes memory use by 6x while keeping 95% of its smarts.
12. [**Ettin-Reranker-1b-V1**](https://huggingface.co/cross-encoder/ettin-reranker-1b-v1) - Delivers speedy relevancy checks locally.
13. [**Command-A-Plus-05-2026-Bf16**](https://huggingface.co/CohereLabs/command-a-plus-05-2026-bf16) - Arrives with 128K context and agentic reasoning.
14. [**Nandi-Mini-600M-Early-Checkpoint**](https://huggingface.co/FrontiersMind/Nandi-Mini-600M-Early-Checkpoint) - Brings 12-language AI to home labs.
15. [**Ring-2.6-1T**](https://huggingface.co/inclusionAI/Ring-2.6-1T) - Brings trillion-parameter reasoning to agentic workflows.
16. [**HRM-Text-1B**](https://huggingface.co/sapientinc/HRM-Text-1B) - Bends time with dual recurrent loops for deep reasoning.
17. [**Nvidia Kimi-K2.6-NVFP4**](https://huggingface.co/nvidia/Kimi-K2.6-NVFP4) - A plug-and-play AI giant optimized for GPUs.
18. [**DeepSeek V4 GGUF**](https://huggingface.co/antirez/deepseek-v4-gguf) - Shrinks the massive DeepSeek V4 for local use.
19. [**Emo**](https://huggingface.co/allenai/emo) - Cuts memory use by 75% with topic-specialized experts.
20. [**NVIDIA-Nemotron-Labs-3-Elastic-30B-A3B-BF16**](https://huggingface.co/nvidia/NVIDIA-Nemotron-Labs-3-Elastic-30B-A3B-BF16) - Unfolds three models in one for flexibility.
21. [**AntAngelMed**](https://huggingface.co/MedAIBase/AntAngelMed) - Deploys a 100B clinical MoE model locally.
22. [**Gemma-4-31B-It-DFlash**](https://huggingface.co/z-lab/gemma-4-31B-it-DFlash) - Drafts speed into your local LLM.
23. [**Leanly_AI**](https://huggingface.co/jackxinning/Leanly_AI) - Arms obesity specialists with empathy backed by health data.
24. [**ZAYA1-8B**](https://huggingface.co/Zyphra/ZAYA1-8B) - Drops a compact reasoning engine for local math and code.
25. [**IBM Granite-4.1-30b**](https://huggingface.co/ibm-granite/granite-4.1-30b) - Empowers private AI agents with multi-tool skills.
26. [**AEON-7 Qwen3.6-27B-AEON-Ultimate-Uncensored-BF16**](https://huggingface.co/AEON-7/Qwen3.6-27B-AEON-Ultimate-Uncensored-BF16) - Unlocks Qwen3.6 with no refusals.
27. [**Ling-2.6-1T**](https://huggingface.co/inclusionAI/Ling-2.6-1T) - Makes trillion parameter AI fast and affordable.
28. [**IBM Granite-4.1-8b**](https://huggingface.co/ibm-granite/granite-4.1-8b) - Advances multilingual chat and tool assistants.
29. [**Hy-MT1.5-1.8B-1.25bit**](https://huggingface.co/AngelSlim/Hy-MT1.5-1.8B-1.25bit) - Puts 33-language translation in your pocket.

**🔀 Multimodal**

1. [**Step-3.7-Flash MoE