9baMelo
## Model Description
This model was trained with the ModelScope AIGC Zone training environment and compute resources.
Model type: LoRA
Base model: krea-community/Krea-2-Turbo
Training code: DiffSynth-Studio
Total training steps: 6000
License: CC BY-NC-SA 4.0
The weight files c1-st1000 ~ c1-st6000 correspond to different training steps. c1-st4000 is recommended.
A LoRA weight of 0.8 is recommended; 1.0 tends to overfit.
## Inference
Install DiffSynth-Studio:
```bash
pip install diffsynth
```
Run inference:
```python
from diffsynth.pipelines.krea2 import Krea2Pipeline, ModelConfig
import torch
pipe = Krea2Pipeline.from_pretrained(
torch_dtype=torch.bfloat16,
device="cuda",
model_configs=[
ModelConfig(model_id="krea-community/Krea-2-Turbo", origin_file_pattern="turbo.safetensors"),
ModelConfig(model_id="Qwen/Qwen3-VL-4B-Instruct", origin_file_pattern="*.safetensors"),
ModelConfig(model_id="Qwen/Qwen-Image", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
],
tokenizer_config=ModelConfig(model_id="Qwen/Qwen3-VL-4B-Instruct", origin_file_pattern=""),
)
pipe.load_lora(pipe.dit, ModelConfig(model_id="toko279/9baMelo", origin_file_pattern="9baMelo_c1-st4000.safetensors"))
prompt = "a cat"
image = pipe(
prompt, seed=0,
num_inference_steps=8, cfg_scale=1, mu=1.15,
)
image.save("image.jpg")
```
## Notice
Krea 2 is licensed under the Krea 2 Community License Agreement. For more information, visit https://krea.ai/krea-2-licensing.
This model is a derivative model fine-tuned from Krea 2 Turbo.