There was an error when uploading the model. It was uploaded twice. You can download either one as they are the same. Complaining about the network of Civitai. It's so slow!
If you'd like to support me and do more : ☕ https://ko-fi.com/vjleoliu ☕commissions open
Model Introduction
A photography LoRA that emulates the retro film style of a "Polaroid". This is the Qwen-image version. Those who need FLUX.1, click here.
Effect comparison
The following comparative tests all use the same parameters, seeds, and prompt words. You can intuitively feel the more prominent effect after using LoRA. (To be honest, it seems that the compliance with the prompt words is higher when LoRA is added, especially in terms of composition.)
Parameter recommendation
Recommended weight: 0.6 - 1.5 (default 0.9)
Recommended size: 768/1024/1536 (default 1024)
Sampling algorithm: 3.0 - 3.5 (default 3.5)
Step size: 20 - 40 (with the acceleration model, it can be reduced to 4 - 8 steps)
Sampler: euler/heun
Cover prompt example:
polaroid style,
Shooting two Asians from a top - down perspective in a narrow and dimly lit indoor space. On the left is a blonde female facing the camera, looking straight ahead. Her hair is pulled back, revealing her forehead. She has delicate makeup, especially with obvious eyeliner and slightly open lips. She is wearing a silver - white sleeveless turtleneck top, and the material looks shiny. She has a small earring in her left ear. There is a wisp of white smoke by her mouth, as if she has just exhaled. The light shines on her cheeks, causing some over - exposure.
On the right is a dark - haired male who slightly raises his head and tilts his body slightly towards the center of the frame. His hair is long, with some strands falling in front of his face. He is wearing black clothes, and he raises his left hand to his mouth. There are tattoos on his fingers, and he is holding a thin white object in his hand, possibly a cigarette. He also has an earring.
The background is a gray wall with stains and signs of wear. There is a white square power socket on the wall, with a black wire connected below it. On the right, there is a curved white pipe running along the wall. The ground is dark in color and a bit blurry, but some debris can be seen, including a metal barrel - like object.
The entire environment appears simple and slightly messy. The lighting is on the cool side, mainly focused on the characters, with the surroundings being rather dim.
🙈🙉🙊🐵OK! That's all
If you like my work, please give it a 👍.
Thank you for reading this far.
English is not my native language. The above is translated by AI. If there is anything inappropriate, welcome to correct me. Finally, I hope you have fun！
👉👉👉My CivitAI homepage👈👈👈
FAQ
Comments (11)
Why are there two files in the file list, and what's the difference between them?
I wrote in the introduction that due to network issues, the same model was uploaded twice. In fact, they are identical.
If you use it with other lora, it works just by giving 0.5 to 0.6 weight. Thank you always for your effort and results
Thank you for your testing and sharing. You have made the application of this LoRA more complete.
Yup, getting some good results at lower weights and omitting the "polaroid style" trigger word as well
@Cedar8043 Hahaha, you noticed it! Yes, the trigger word is just an extra safeguard. Qwen-image can be quite stubborn sometimes.
I really like this in conjunction with the smartphone-snapshot v4 LoRA. Creates a really interesting analog aesthetic.
It's true that many people like to use smartphones with some of my realistic models. Although I don't know what kind of aesthetic works you have generated, this makes me very curious about smartphones.
this lora is the best thanks for making it, by any chance were you going to train one for the new z-Image Turbo Model?