Instructions to use perpetual3x/ROC_Vintage_Poster with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use perpetual3x/ROC_Vintage_Poster with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("Qwen/Qwen-Image", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("perpetual3x/ROC_Vintage_Poster") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Inference
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
ROC_Vintage_Poster
A vintage poster style LoRA inspired by Republic of China (ROC) 1930s commercial posters, Shanghai calendar posters, and classic lithographic printing aesthetics.
This LoRA recreates the visual language of early 20th century Chinese posters:
- vintage commercial illustration
- hand-painted lithograph style
- old paper texture
- Republican China atmosphere
- Shanghai retro advertising aesthetics
The goal is not historical reconstruction, but a creative exploration of an imagined visual world from the 1930s.
โจ Features
- 1930s Republic of China vintage poster style
- Shanghai calendar poster aesthetics
- Classic commercial advertisement composition
- Hand-painted illustration feeling
- Lithograph printing texture
- Retro typography layout
- Aged paper appearance
๐ฅ๏ธ Training Details
Hardware
Training was performed on:
- GPU: NVIDIA RTX PRO 6000 Blackwell 96GB
Configuration
- Base Model: Qwen Image Text-to-Image (original version)
- LoRA Rank: 32
- Batch Size: 8
- Steps: 3000
โ ๏ธ Important: Base Model Compatibility
This LoRA was trained with:
Qwen Image Text-to-Image (first release)
It was NOT trained with Qwen Image 2512.
Qwen Image versions have significant differences in architecture and behavior.
Using this LoRA with other versions may produce different results.
For the best experience, please use the original Qwen Image Text-to-Image model.
๐ซ About NSFW
This LoRA was not trained for NSFW generation.
The purpose of this project is:
- vintage poster style
- historical-inspired artwork
- creative image generation
If you are looking for NSFW-specific LoRAs, please explore other third-party projects.
๐ฐ๏ธ A Small Historical Note
A small problem with generating images from the 1930s:
Most of us did not experience Shanghai in the 1930s firsthand.
(I haven't either.)
So when you ask:
"Generate a realistic street scene in Shanghai in 1935"
there is no single correct answer.
The result depends on:
- your imagination
- your prompt
- the model's interpretation
Feel free to edit prompts, experiment with different descriptions, or ask an AI assistant for help.
After all, imagining the past is also a creative journey.
๐ผ๏ธ Examples
The following images were generated with this LoRA.
The prompts below are the original prompts used for generation.
Feel free to copy, modify, and experiment with them.
Example 1
Prompt
Draw zl_vintage_v1, ROC calendar poster illustration, a beautiful Chinese woman in 1930s Shanghai, wearing an elegant silk qipao, standing in a luxurious art deco interior, soft smile, graceful pose, vintage Chinese advertising poster, Republican era Shanghai calendar poster, hand painted illustration, lithograph print texture, warm colors, delicate facial features, elegant feminine beauty, retro commercial advertisement, fine details, aged paper texture.
Example 2
Prompt
Draw zl_vintage_v1, ROC calendar poster illustration, 1930s Chinese beauty advertisement poster, a glamorous Chinese woman wearing a traditional qipao, holding a perfume bottle, Shanghai old fashion style, vintage cosmetics advertisement, Republic of China era, hand painted poster art, realistic face with illustration style, elegant hairstyle, silk dress, art deco background, soft lighting, muted colors, old paper grain, authentic 1930s poster.
Example 3
Prompt
Draw zl_vintage_v1, ROC calendar poster illustration, a beautiful young Chinese woman in 1930s Shanghai, wearing a Republican era school uniform, short bob hairstyle, holding books, standing on an old Shanghai street, vintage Chinese poster illustration, calendar girl poster, hand painted style, lithography texture, nostalgic atmosphere, elegant and innocent expression, warm vintage color palette.
Example 4
Prompt
Draw zl_vintage_v1, ROC calendar poster illustration, elegant Chinese lady from 1930s Shanghai, wearing luxurious silk qipao and pearl necklace, sitting beside an antique radio, art deco apartment, vintage Republican China advertisement poster, beautiful woman portrait, hand painted commercial illustration, Shanghai golden age style, detailed face, soft brush strokes, aged paper texture.
Example 5
Prompt
Draw zl_vintage_v1, ROC calendar poster illustration, 1930s Chinese movie star poster, glamorous Shanghai actress wearing elegant cheongsam, dramatic pose, classic cinema advertisement, Republic of China era film poster, hand painted lithograph, vintage typography area, art deco design, soft glamorous lighting, beautiful Chinese face, nostalgic old Shanghai atmosphere.
๐ Inference Example
- This example loads the full Qwen-Image model in BF16 without quantization, CPU offloading, or inference acceleration. A high-memory CUDA GPU is required.
- An 80 GB GPU is recommended for reliable inference; GPUs with 48 GB of VRAM may work depending on the software environment, image resolution, and available system memory. Most consumer GPUs will require memory-saving techniques that are not included in this baseline example.
- This is a baseline example for reproducing the LoRA with the original, unquantized Qwen-Image pipeline. Memory-saving and accelerated configurations are intentionally not included.
This LoRA was trained for the original Qwen/Qwen-Image model.
Trigger word: zl_vintage_v1
Recommended prompt prefix:zl_vintage_v1, ROC calendar poster illustration
Installation
# Install a CUDA-compatible version of PyTorch first:
# https://pytorch.org/get-started/locally/
pip install -U diffusers transformers accelerate safetensors
Python
import torch
from diffusers import QwenImagePipeline
base_model_id = "Qwen/Qwen-Image"
lora_model_id = "perpetual3x/ROC_Vintage_Poster"
# Load the original Qwen-Image model without quantization or acceleration adapters.
pipe = QwenImagePipeline.from_pretrained(
base_model_id,
torch_dtype=torch.bfloat16,
)
# Load the ROC Vintage Poster LoRA from Hugging Face.
pipe.load_lora_weights(
lora_model_id,
weight_name="rocvintage.safetensors",
)
pipe = pipe.to("cuda")
prompt = (
"Draw zl_vintage_v1, ROC calendar poster illustration, a beautiful Chinese woman in 1930s Shanghai, wearing an elegant silk qipao, standing in a luxurious art deco interior, soft smile, graceful pose, vintage Chinese advertising poster, Republican era Shanghai calendar poster, hand painted illustration, lithograph print texture, warm colors, delicate facial features, elegant feminine beauty, retro commercial advertisement, fine details, aged paper texture."
)
negative_prompt = (
"low quality, blurry, distorted anatomy, extra fingers, "
"modern clothing, modern photography, watermark"
)
generator = torch.Generator(device="cuda").manual_seed(42)
image = pipe(
prompt=prompt,
negative_prompt=negative_prompt,
width=1328,
height=1328,
num_inference_steps=50,
true_cfg_scale=4.0,
generator=generator,
).images[0]
image.save("roc_vintage_poster.png")
โ ๏ธ License & Copyright Notice
This LoRA may be used for personal and non-commercial purposes only.
You may:
- Use this LoRA to create personal artwork
- Experiment with AI image generation
- Share generated images with proper attribution
You may not:
- Use this LoRA for commercial purposes
- Sell generated content as a commercial product
- Redistribute, repackage, or sell this LoRA
- Include this LoRA in paid services or products
For commercial licensing inquiries, please contact the author.
- Downloads last month
- 25
Model tree for perpetual3x/ROC_Vintage_Poster
Base model
Qwen/Qwen-Image




