Modelpublic

furaidosu/tostiadv-pilot-qwen-image-2512

sourceHugging Faceapache-2.0updated 2mo agoView on Hugging Face

0likes

Model Card

tostiadv-pilot-qwen-image-2512

LoRA para Qwen-Image-2512, entrenado originalmente en ModelScope (tostiok/tostiadv_pilot_qwen_image_2512) con DiffSynth-Studio.

Trigger word: tostiadv_style

—Dataset: 15 imágenes
—Total de pasos de entrenamiento: 4000

Checkpoints

Step	Archivo	Muestra
1000	`tostiadv_pilot_qwen_image_2512_c1-st1000.safetensors`	<img src="https://huggingface.co/furaidosu/tostiadv-pilot-qwen-image-2512/resolve/main/images/st1000.png" width="320" alt="step 1000">
2000	`tostiadv_pilot_qwen_image_2512_c1-st2000.safetensors`	<img src="https://huggingface.co/furaidosu/tostiadv-pilot-qwen-image-2512/resolve/main/images/st2000.png" width="320" alt="step 2000">
3000	`tostiadv_pilot_qwen_image_2512_c1-st3000.safetensors`	<img src="https://huggingface.co/furaidosu/tostiadv-pilot-qwen-image-2512/resolve/main/images/st3000.png" width="320" alt="step 3000">
4000	`tostiadv_pilot_qwen_image_2512_c1-st4000.safetensors`	<img src="https://huggingface.co/furaidosu/tostiadv-pilot-qwen-image-2512/resolve/main/images/st4000.png" width="320" alt="step 4000">

Uso con diffusers

python

from diffusers import DiffusionPipeline
import torch

pipe = DiffusionPipeline.from_pretrained("Qwen/Qwen-Image-2512", torch_dtype=torch.bfloat16)
pipe.load_lora_weights(
    "furaidosu/tostiadv-pilot-qwen-image-2512",
    weight_name="tostiadv_pilot_qwen_image_2512_c1-st4000.safetensors",
)
pipe.to("cuda")

image = pipe("tostiadv_style, tu prompt acá").images[0]