---
license: apache-2.0
---
# Templates-Inpainting (FLUX.2-klein-base-4B)

This model is one of the open-source Diffusion Templates series models from [DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio). Specifically designed for inpainting tasks, it accepts an original image and a mask image, then generates new content within the masked region based on natural language prompts, seamlessly blending with the surrounding unmasked background.

* Open-source code: [DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio)
* Technical report: [arXiv](https://arxiv.org/abs/2604.24351)
* Project page: [GitHub](https://modelscope.github.io/diffusion-templates-web/)
* Documentation: [English Version](https://diffsynth-studio-doc.readthedocs.io/en/latest/Diffusion_Templates/Introducing_Diffusion_Templates.html)、[中文版](https://diffsynth-studio-doc.readthedocs.io/zh-cn/latest/Diffusion_Templates/Introducing_Diffusion_Templates.html)
* Online demo: [ModelScope](https://modelscope.cn/studios/DiffSynth-Studio/Diffusion-Templates)
* Models: [ModelScope](https://modelscope.cn/collections/DiffSynth-Studio/KleinBase4B-Templates)、[ModelScope International](https://modelscope.ai/collections/DiffSynth-Studio/KleinBase4B-Templates)、[HuggingFace](https://huggingface.co/collections/DiffSynth-Studio/kleinbase4b-templates)
* Datasets: [ModelScope](https://modelscope.cn/collections/DiffSynth-Studio/ImagePulseV2)、[ModelScope International](https://modelscope.cn/collections/DiffSynth-Studio/ImagePulseV2)、[HuggingFace](https://huggingface.co/collections/DiffSynth-Studio/imagepulsev2)

## Results

| Reference | Prompt | Mask | Generated |
|:---:|:---|:---:|:---:|
| ![](./assets/cat_base.jpg) | An orange cat is sitting on a stone. | ![](./assets/cat_mask_1.jpg) | ![](./assets/cat_Inpaint_1.jpg) |
| ![](./assets/cat_base.jpg) | A cat wearing sunglasses is sitting on a stone. | ![](./assets/cat_mask_2.jpg) | ![](./assets/cat_Inpaint_2.jpg) |

| Reference | Prompt | Mask | Generated |
|:---:|:---|:---:|:---:|
| ![](./assets/girl_base.jpg) | A beautiful young woman wearing a woven straw hat with a ribbon standing in a sunflower field. | ![](./assets/girl_mask_1.jpg) | ![](./assets/girl_Inpaint_1.jpg) |
| ![](./assets/girl_base.jpg) | A beautiful young woman wearing an elegant white dress standing in a glowing sunflower field. | ![](./assets/girl_mask_2.jpg) | ![](./assets/girl_Inpaint_2.jpg) |

| Reference | Prompt | Mask | Generated |
|:---:|:---|:---:|:---:|
| ![](./assets/room_base.jpg) | A sleek glass vase with a single blooming white lily and an open minimalist art book resting on the circular white marble coffee table. | ![](./assets/room1.jpg) | ![](./assets/room_Inpaint_1.jpg) |
| ![](./assets/room_base.jpg) | A large, minimalist flower painting hanging on the clean off-white wall above the sofa, soft shadows. | ![](./assets/room2.jpg) | ![](./assets/room_Inpaint_2.jpg) |

## Inference Code

* Install [DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio)

```
git clone https://github.com/modelscope/DiffSynth-Studio.git
cd DiffSynth-Studio
pip install -e .
```

* Direct inference (requires 40GB GPU memory)

```python
from diffsynth.diffusion.template import TemplatePipeline
from diffsynth.pipelines.flux2_image import Flux2ImagePipeline, ModelConfig
import torch
from modelscope import dataset_snapshot_download
from PIL import Image
```

```python
pipe = Flux2ImagePipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[
        ModelConfig(model_id="black-forest-labs/FLUX.2-klein-base-4B", origin_file_pattern="transformer/*.safetensors"),
        ModelConfig(model_id="black-forest-labs/FLUX.2-klein-4B", origin_file_pattern="text_encoder/*.safetensors"),
        ModelConfig(model_id="black-forest-labs/FLUX.2-klein-4B", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
    ],
    tokenizer_config=ModelConfig(model_id="black-forest-labs/FLUX.2-klein-4B", origin_file_pattern="tokenizer/"),
)
template = TemplatePipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[ModelConfig(model_id="DiffSynth-Studio/Template-KleinBase4B-Inpaint")],
)
dataset_snapshot_download(
    "DiffSynth-Studio/examples_in_diffsynth",
    allow_file_pattern=["templates/*"],
    local_dir="data/examples",
)
image = template(
    pipe,
    prompt="An orange cat is sitting on a stone.",
    seed=0, cfg_scale=4, num_inference_steps=50,
    template_inputs = [{
        "image": Image.open("data/examples/templates/image_reference.jpg"),
        "mask": Image.open("data/examples/templates/image_mask_1.jpg"),
        "force_inpaint": True,
    }],
    negative_template_inputs = [{
        "image": Image.open("data/examples/templates/image_reference.jpg"),
        "mask": Image.open("data/examples/templates/image_mask_1.jpg"),
    }],
)
image.save("image_Inpaint_1.jpg")
image = template(
    pipe,
    prompt="A cat wearing sunglasses is sitting on a stone.",
    seed=0, cfg_scale=4, num_inference_steps=50,
    template_inputs = [{
        "image": Image.open("data/examples/templates/image_reference.jpg"),
        "mask": Image.open("data/examples/templates/image_mask_2.jpg"),
    }],
    negative_template_inputs = [{
        "image": Image.open("data/examples/templates/image_reference.jpg"),
        "mask": Image.open("data/examples/templates/image_mask_2.jpg"),
    }],
)
image.save("image_Inpaint_2.jpg")
```

* Enable lazy loading and memory management, requires 24G GPU memory

```python
from diffsynth.diffusion.template import TemplatePipeline
from diffsynth.pipelines.flux2_image import Flux2ImagePipeline, ModelConfig
import torch
from modelscope import dataset_snapshot_download
from PIL import Image
```

```python
vram_config = {
    "offload_dtype": "disk",
    "offload_device": "disk",
    "onload_dtype": torch.float8_e4m3fn,
    "onload_device": "cpu",
    "preparing_dtype": torch.float8_e4m3fn,
    "preparing_device": "cuda",
    "computation_dtype": torch.bfloat16,
    "computation_device": "cuda",
}
pipe = Flux2ImagePipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[
        ModelConfig(model_id="black-forest-labs/FLUX.2-klein-base-4B", origin_file_pattern="transformer/*.safetensors", **vram_config),
        ModelConfig(model_id="black-forest-labs/FLUX.2-klein-4B", origin_file_pattern="text_encoder/*.safetensors", **vram_config),
        ModelConfig(model_id="black-forest-labs/FLUX.2-klein-4B", origin_file_pattern="vae/diffusion_pytorch_model.safetensors"),
    ],
    tokenizer_config=ModelConfig(model_id="black-forest-labs/FLUX.2-klein-4B", origin_file_pattern="tokenizer/"),
    vram_limit=torch.cuda.mem_get_info("cuda")[1] / (1024 ** 3) - 0.5,
)
template = TemplatePipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[ModelConfig(model_id="DiffSynth-Studio/Template-KleinBase4B-Inpaint")],
    lazy_loading=True,
)
dataset_snapshot_download(
    "DiffSynth-Studio/examples_in_diffsynth",
    allow_file_pattern=["templates/*"],
    local_dir="data/examples",
)
image = template(
    pipe,
    prompt="An orange cat is sitting on a stone.",
    seed=0, cfg_scale=4, num_inference_steps=50,
    template_inputs=[{
        "image": Image.open("data/examples/templates/image_reference.jpg"),
        "mask": Image.open("data/examples/templates/image_mask_1.jpg"),
        "force_inpaint": True,
    }],
    negative_template_inputs=[{
        "image": Image.open("data/examples/templates/image_reference.jpg"),
        "mask": Image.open("data/examples/templates/image_mask_1.jpg"),
    }],
)
image.save("image_Inpaint_1.jpg")
image = template(
    pipe,
    prompt="A cat wearing sunglasses is sitting on a stone.",
    seed=0, cfg_scale=4, num_inference_steps=50,
    template_inputs=[{
        "image": Image.open("data/examples/templates/image_reference.jpg"),
        "mask": Image.open("data/examples/templates/image_mask_2.jpg"),
    }],
    negative_template_inputs=[{
        "image": Image.open("data/examples/templates/image_reference.jpg"),
        "mask": Image.open("data/examples/templates/image_mask_2.jpg"),
    }],
)
image.save("image_Inpaint_2.jpg")
```

## Training Code

After installing DiffSynth-Studio, use the following script to start training. For more information, please refer to the [DiffSynth-Studio Documentation](https://diffsynth-studio-doc.readthedocs.io/zh-cn/latest/).

```shell
modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "flux2/Template-KleinBase4B-Inpaint/*" --local_dir ./data/diffsynth_example_dataset

accelerate launch examples/flux2/model_training/train.py \
  --dataset_base_path data/diffsynth_example_dataset/flux2/Template-KleinBase4B-Inpaint \
  --dataset_metadata_path data/diffsynth_example_dataset/flux2/Template-KleinBase4B-Inpaint/metadata.jsonl \
  --extra_inputs "template_inputs" \
  --max_pixels 1048576 \
  --dataset_repeat 50 \
  --model_id_with_origin_paths "black-forest-labs/FLUX.2-klein-4B:text_encoder/*.safetensors,black-forest-labs/FLUX.2-klein-base-4B:transformer/*.safetensors,black-forest-labs/FLUX.2-klein-4B:vae/diffusion_pytorch_model.safetensors" \
  --template_model_id_or_path "DiffSynth-Studio/Template-KleinBase4B-Inpaint:" \
  --tokenizer_path "black-forest-labs/FLUX.2-klein-4B:tokenizer/" \
  --learning_rate 1e-4 \
  --num_epochs 2 \
  --remove_prefix_in_ckpt "pipe.template_model." \
  --output_path "./models/train/Template-KleinBase4B-Inpaint_full" \
  --trainable_models "template_model" \
  --use_gradient_checkpointing \
  --find_unused_parameters
```