Vision XL - LCM Edition

CHECKPOINT
Original


Updated:

626

🔎 What is LCM and how does it work?

LCM stands for Latent Consistency Model, which is a novel technique for accelerating text-to-image generation tasks. LCM are distilled from pre-trained latent diffusion models (LDM), which are powerful generative models that can produce high-quality images from text prompts. LCM learn to mimic the behavior of LDM, but with much fewer inference steps and lower computational cost. LCM achieve this by using a latent consistency loss, which ensures that the latent codes of LCM are close to those of LDM at each step. LCM also use a LoRA distillation method, which reduces the memory consumption and improves the image quality.

It’s like the old one but better one.

Recommended parameters for both versions are almost the same as before, with a few improvements and changes.

Achieving super high-quality images with super low steps and CFG Scale is actually really challenging. I performed countless experiments, and each time, I didn’t achieve what i wanted.

So, I decided to improve the previous version. Now, this version is pretty good. With in 8-10 steps and 4-5 CFG Scale, you will get results comparable to the normal SDXL with 30-40 steps.

If you like my work, then please share your feedback in the comment section and don't forget to share your work. I'm excited to see them.

Want to support me, you can check my Buffet plan. Your support means a lot to me and it motivates me to keep improving.

✅Recommended Parameters✅

  • VAE : SDXL VAE

  • Sampling Method : DPM++ 2M Karras | You can use different sampling methods, but personally, I'm a fan of DPM++ 2M Karras. So, if you want to try another sampler, please give it a try and share your experience with me

  • Sampling Steps : 8➡️16 | safe starting 8/10

  • CFG Scale : 3➡️6 | 4/5 safe starting

  • Clip Skip : 1

  • Upscaler : 4x-UltraSharp

  • Hires Steps : 4➡️10

  • Denoising Strength : 0.1➡️0.5

Join my community, Share your feedback, learn, and have fun with us! 😊

Discord➡️https://discord.gg/QQKd7bu97P

User Guide 📢

I'm here to help you create awesome visuals! Remember, I can't be responsible if your content breaks any laws or community guidelines ❌

Let's collaborate creatively, follow the rules, and make something amazing together 🎨

📜 If you have questions, feel free to ask anytime! 🌟

Version Detail

SDXL 1.0
Hyperreal Surrealist SDXL LCM base model by fusing it with LCM SDXL Lora and making some changes. Although the model was ready within a day, testing took longer than expected. After testing, I made some changes, and finally, I achieved the desired results. Please Note: As this model is still under development, you may encounter occasional imperfections or inconsistencies in the generated images. We are continuously working on addressing these issues and improving the model's performance. Feedback and Contribution: We welcome your feedback and suggestions as we strive to refine this model. Please share your experiences and provide any insights that could help us enhance its capabilities.

Project Permissions

Reprinting is strictly prohibited

    Use Permissions

  • Use in TENSOR Online

  • As a online training base model on TENSOR

  • Use without crediting me

  • Share merges of this model

  • Use different permissions on merges

    Commercial Use

  • Sell generated contents

  • Use on generation services

  • Sell this model or merges

Comments

Related Posts

Describe the image you want to generate, then press Enter to send.