Text-to-Image
Safetensors
PyTorch
moss-image
moss-image-t2i
moss-image-edit
moss-image-editing
moss-image-generation
moss-image-model
moss-ai
comfy-org
text-to-image-generation
image-to-image
image-to-image-generation
image-editing
ai-image-editing
ai-image-generation
image-transformation
image-modification
image-manipulation
image-reimagining
image-enhancement
generative-image-editing
instruction-based-image-editing
text-guided-image-editing
prompt-based-image-editing
natural-language-image-editing
image-instruction
instruction-following
reference-image
source-image
image-conditioning
image-understanding
character-consistency
pose-preservation
background-preservation
image-generation-model
ai-image-generator
generative-ai
generative-art
ai-art
digital-art
creative-ai
diffusion
diffusion-model
diffusion-models
diffusion-transformer
dit
flow-matching
image-synthesis
text-to-image-model
image-to-image-model
base-model
instruction-tuned
instruction-tuned-model
sft
supervised-fine-tuning
foundation-model
foundation-models
qwen
qwen3
qwen3-5
qwen3-5-9b
qwen-text-encoder
text-encoder
language-model
multimodal-model
flux
flux2
flux2-vae
vae
image-vae
latent-space
lora
lora-model
edit-lora
image-edit-lora
lora-support
lora-integration
lora-loading
custom-lora
comfyui
comfyui-workflow
comfyui-custom-nodes
comfyui-api
comfyui-image-generation
comfyui-image-editing
node-based-workflow
workflow
xformers
memory-efficient-attention
skip-layer-guidance
slg
sampler
sampling
scheduler
cfg
seed-control
image-resolution
image-resizing
prompt-engineering
positive-prompt
negative-prompt
image-prompting
image-processing
computer-vision
deep-learning
machine-learning
artificial-intelligence
neural-networks
generative-models
python
python-project
model-inference
inference
google-colab
colab-notebook
jupyter-notebook
google-colab-gpu
nvidia-t4
gpu-acceleration
cuda
cloud-gpu
aria2
model-downloader
ai-workflow
ai-tools
open-source-ai
image-generation-tool
image-editing-tool
ai-photo-editor
photo-editing
image-creation
image-editing-tutorial
image-generation-tutorial
tutorial
beginner-friendly
no-code
low-code
interactive-ai
creative-workflow
machine-learning-project
ai-project

โšก MOSS-Image T2I & Edit in Google Colab

This repository contains a simple, automated Google Colab notebook for running the MOSS-Image AI model. Built on a ComfyUI backend, it allows you to generate stunning images from text (Text-to-Image) and perform precise edits on existing photos (Image-to-Image), completely free on a Google Colab T4 GPU.

๐ŸŽฅ Watch the Tutorial: How to Use

๐Ÿš€ Run in Colab: Open Google Colab Notebook

Open In Colab Get Pro


โœจ Features Supported in this Notebook

This notebook automates the MOSS-Image workflow into 4 simple cells:

  1. โš™๏ธ Initialize Core Environment: Installs the ComfyUI backend and required dependencies like xformers.
  2. ๐Ÿ“ฅ High-Speed Asset Downloader: Uses Aria2c to rapidly fetch the MOSS-Image checkpoints (SFT and Base), the Qwen3.5 9B text encoder, and the Flux2-VAE. Automatically downloads the required Edit LoRA and supports custom user LoRAs.
  3. ๐ŸŽจ Text-to-Image Generation:
    • Generate images from scratch using detailed prompts.
    • Choose between the Instruction-Tuned (SFT) or Raw (Base) models.
    • Features Advanced Sampler Settings and 'Skip Layer Guidance' (SLG) for enhanced realism.
  4. ๐Ÿ–Œ๏ธ MOSS-Image Edit:
    • Upload an existing source image and provide a text instruction (e.g., "Change the shirt to a black leather jacket").
    • The dedicated Edit LoRA redraws the image with your changes while perfectly preserving the character, pose, and background.

๐Ÿ› ๏ธ How to Use

  1. Click the "Open in Colab" badge above.
  2. Go to Runtime > Change runtime type and select the T4 GPU.
  3. Run Cell 1 to initialize the ComfyUI core environment.
  4. Go to Cell 2, choose which models you want to download (SFT, Base, or Both), and hit Play. You can also upload a custom LoRA here if you have one.
  5. To generate an image: Go to Cell 3, type your description in the PROMPT box, adjust your width/height, and click Play.
  6. To edit an image: Go to Cell 4, type your edit instructions in the EDIT_INSTRUCTION box, and click Play. You will be prompted to upload your source image, and the AI will apply your changes.

๐Ÿค Credits

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support