Glitch-Img-1.0: RGBA Image Generation & Editing with Qwen-Image-2.1

Developer Islam Al-Nashar re-uploaded Qwen's Qwen-Image-2.1 model to Hugging Face as ISLAM-PO/Glitch-Img-1.0, keeping all weights and configs unchanged while adding documentation and a demo Space. The model supports text-to-image generation up to 2048x2048, inpainting-style editing, native RGBA transparency output, and up to 10 reference images for subject consistency, and is released under the non-commercial Qwen Research License. Al-Nashar said he uses it as the visual layer of an Arabic-first AI stack alongside his text model MasryGPT.

Try it: huggingface.co/ISLAM-PO/Glitch-Img-1.0 https://huggingface.co/ISLAM-PO/Glitch-Img-1.0 · image-gen demo Space https://huggingface.co/spaces/ISLAM-PO/image-gen-demo Image generation has moved past "pretty pictures" toward controllable, editable, transparent assets . Glitch-Img-1.0 is my re-upload of Qwen/Qwen-Image-2.1 under the name ISLAM-PO/Glitch-Img-1.0 , aimed at making a strong unified text-to-image + editing model easy to discover and use from a single, well-documented repository. All weights and configs are unchanged from the original; only the README and repo IDs were updated, and the model is Built with Qwen under the Qwen Research License non-commercial . This report covers what the model does, its key capabilities, a quick-start, memory tips, and licensing notes. Qwen/Qwen-Image-2.1 text-to-image diffusers: QwenImage21Pipeline Four headline improvements define the Qwen-Image-2.1 family this build inherits: | Capability | Details | |---|---| | Text-to-image | Prompts up to 2048×2048 and beyond; strong text rendering | | Image editing | Inpainting-style edits, background swaps, style transfer | | RGBA transparency | Native alpha channel generation with a documented prompt format | | Multi-reference | Up to 10 reference images for identity/subject consistency | | Local edits | Circles, painted annotations, separate masks | | Aspect ratios | 1:1, 4:3, 3:4, 3:2, 2:3, 16:9, 9:16 presets | Install: pip install torch =2.4.0 pip install transformers =5.17 pip install git+https://github.com/huggingface/diffusers pip install accelerate pillow Text-to-image: python import torch from diffusers import QwenImage21Pipeline pipe = QwenImage21Pipeline.from pretrained "ISLAM-PO/Glitch-Img-1.0", torch dtype=torch.bfloat16 .to "cuda" image = pipe prompt='A neon shop sign that reads "GLITCH IMG 1.0", rainy night, reflections on wet pavement', width=2048, height=2048, num inference steps=40, generator=torch.Generator "cuda" .manual seed 42 , .images 0 image.save "t2i example.png" Image editing: python import torch from PIL import Image from diffusers import QwenImage21Pipeline pipe = QwenImage21Pipeline.from pretrained "ISLAM-PO/Glitch-Img-1.0", torch dtype=torch.bfloat16 .to "cuda" input image = Image.open "input.png" image = pipe prompt="Change the background to a sunset beach", image=input image, num inference steps=40, generator=torch.Generator "cuda" .manual seed 42 , .images 0 image.save "edit example.png" Transparent RGBA generation — use the recommended prompt format: image = pipe prompt="This is an RGBA image with transparency. A cute cartoon dragon sticker. " "The image has alpha channel and the background is transparent.", width=2048, height=2048, num inference steps=40, generator=torch.Generator "cuda" .manual seed 42 , .images 0 image.save "transparent example.png" Aspect ratios: aspect ratios = { "1:1": 2048, 2048 , "4:3": 2400, 1792 , "3:4": 1792, 2400 , "3:2": 2528, 1696 , "2:3": 1696, 2528 , "16:9": 2752, 1536 , "9:16": 1536, 2752 , } pipe = QwenImage21Pipeline.from pretrained "ISLAM-PO/Glitch-Img-1.0", torch dtype=torch.bfloat16 pipe.enable model cpu offload This offloads model parts to CPU between steps, letting you run on smaller GPUs at the cost of speed. I use Glitch-Img as the visual layer behind my portfolio and product work: It complements MasryGPT text as the two halves of an Arabic-first AI product stack: one generates the words, the other the visuals. model-business@notice.qwencloud.com . NOTICE and LICENSE in the repo for required attribution. Author: Islam Al-Nashar — Full-Stack Developer & AI builder. GitHub https://github.com/aslamalkarywk7 · Hugging Face https://huggingface.co/ISLAM-PO · Portfolio https://aslamalkarywk7.github.io/aslamalkarywk7/