#As requested: AI models that can mimic "Muse" but for free (Will take some extra steps)

1 messages · Page 1 of 1 (latest)

clever yew
#

Note for (Texture generation, Sprite generation and 2D Enhancers its recommended to use Comfy-Ui as its the most widely adopted inference engine for image generation)

Comfy-ui:
https://github.com/comfyanonymous/ComfyUI

Texture generation:

  1. https://huggingface.co/dream-textures/texture-diffusion
  2. https://huggingface.co/sshh12/sdxl-lora-planet-textures
  3. https://huggingface.co/gokaygokay/Flux-Seamless-Texture-LoRA (This one is a lora adapter so you need to merge it with the base model for it to be usable)

Chat/Code generation: (My personal recommendations - Use LM studio for inference) Based on vram size (Ram size on gpu) Note: the higher up on Vram you go the smarter the models become:
6gb: https://huggingface.co/bartowski/open-r1_OlympicCoder-7B-GGUF/blob/main/open-r1_OlympicCoder-7B-Q4_K_L.gguf
12gb: https://huggingface.co/bartowski/phi-4-GGUF/blob/main/phi-4-Q5_K_L.gguf
16gb: https://huggingface.co/bartowski/RekaAI_reka-flash-3-GGUF/blob/main/RekaAI_reka-flash-3-Q4_K_L.gguf
24gb: https://huggingface.co/bartowski/Qwen_QwQ-32B-GGUF/blob/main/Qwen_QwQ-32B-Q4_K_L.gguf
32gb: https://huggingface.co/bartowski/nvidia_Llama-3_3-Nemotron-Super-49B-v1-GGUF/blob/main/nvidia_Llama-3_3-Nemotron-Super-49B-v1-Q4_K_S.gguf
48gb: https://huggingface.co/bartowski/nvidia_Llama-3_3-Nemotron-Super-49B-v1-GGUF/blob/main/nvidia_Llama-3_3-Nemotron-Super-49B-v1-Q5_K_L.gguf

LM studio (For using models above):
https://lmstudio.ai

Motion Animation: 😢 This doesnt seem to exist in open source yet sadly

Behavior: Likely easier to create through code with the code generation AI rather than try and implement a way for the AI to create literal behavior trees. (No behavior specific AI exist)

#

Sprite generation:

  1. https://huggingface.co/Onodofthenorth/SD_PixelArt_SpriteSheet_Generator
  2. https://huggingface.co/ana-tamais/2d-art-sprites
  3. https://huggingface.co/Verah/spine-sprite-lora-v0.1 (This one is a lora adapter so you need to merge it with the base model for it to be usable)

2D Enhancers:

Upscale: https://huggingface.co/black-forest-labs/FLUX.1-dev (this can also be used to generate images)
Recolor: https://huggingface.co/briaai/BRIA-2.3-ControlNet-Recoloring
Pixelate: https://huggingface.co/ntc-ai/SDXL-LoRA-slider.pixelated (This one is a lora adapter so you need to merge it with the base model for it to be usable)
Remove Background: https://huggingface.co/briaai/RMBG-1.4

placid mural
#

Nice resource, thanks! i will definitely be trying these shortly.
For those who don't like AI, that is fine. no need to troll a resource thread for those who Are interested though.

clever yew
south swan
split elbow
#

can I please ask how do you find using muse? Is it worth the learning curve to try out and experiment with? I am already subscribed to chat gpt and am thinking about cursor so its potentially not feasible for me to pay the subscription in the long term. However it seems like I can use muse to learn about AI processes in Unity and then branch out to using the sort of apps you describe? Thank you very much for your insights! 🙏

#

I have started using it before but the delay period in answering my messages meant I switched back to chat gpt for my questions.

#

I am wondering if someone has used it recently and so could provide insights