Spaces:

---

selfit-camera

/

Omni-Image-Editor

like 2.08k

Running on CPU Upgrade

There's no open source model here. The model is closed-source and deliberately obfuscated. This space is designed to drive traffic towards their paid service, which doesn't identify the model(s) used or advance ML research in any meaningful way. This is self-promo and spam

Here is the Comfyui workflow they use, simply copy/download it and put it in Comfyui. Not sure why they were so hell bent on keeping this private, claiming they spend so much money on RnD and shit when they just use trained models from someone else, well probably so people go their website and buy a subscription model but even then why not just tell what you use for the people who have the computers to run it themselves lmao, there are still plenty of people who would have payed money because their pc's just arent powerfull enough

Workflow (copy/dragNdrop into comfyui):
https://pastebin.com/qzrhqENx

and here is the model they use:
https://huggingface.co/Phr00t/Qwen-Image-Edit-Rapid-AIO/tree/main/v19

so much about them "training the model themselves"

btw you can check yourself by just putting the output images in a hexeditor or comfyui itself:
{"1": {"inputs": {"ckpt_name": "v19/Qwen-Rapid-AIO-NSFW-v19.safetensors"}, "class_type": "CheckpointLoaderSimple", "_meta": {"title": "REDACTED"}}, "2": {"inputs": {"seed": REDACTED, "steps": 4, "cfg": 1.0, "sampler_name": "lcm", "scheduler": "beta", "denoise": 1.0, "model": ["1", 0], "positive": ["3", 0], "negative": ["4", 0], "latent_image": ["9", 0]}, "class_type": "KSampler", "_meta": {"title": "REDACTED"}}, "3": {"inputs": {"prompt": "REDACTED", "target_size": 896, "clip": ["1", 1], "vae": ["1", 2], "image1": ["7", 0]}, "class_type": "TextEncodeQwenImageEditPlus", "_meta": {"title": "TextEncodeQwenImageEditPlus Input Prompt"}}, "4": {"inputs": {"prompt": "", "target_size": 800, "clip": ["1", 1], "vae": ["1", 2]}, "class_type": "TextEncodeQwenImageEditPlus", "_meta": {"title": "TextEncodeQwenImageEditPlus Negative (leave blank)"}}, "5": {"inputs": {"samples": ["2", 0], "vae": ["1", 2]}, "class_type": "VAEDecode", "_meta": {"title": "REDACTED"}}, "6": {"inputs": {"images": ["5", 0]}, "class_type": "PreviewImage", "_meta": {"title": "REDACTED"}}, "7": {"inputs": {"image": "REDACTED"}, "class_type": "LoadImage", "_meta": {"title": "Optional Input Image"}, "is_changed": ["REDACTED"]}, "9": {"inputs": {"width": 984, "height": 648, "batch_size": 1}, "class_type": "EmptyLatentImage", "_meta": {"title": "Final Image Size"}}}