使用 `distilabel` 进行图像生成¶

使用 distilabel 创建合成图像。

此示例展示了如何使用 distilabel 生成图像数据，可以使用 InferenceEndpointsImageGeneration 或 OpenAIImageGeneration，这要归功于 ImageGeneration 任务。

Inference Endpoints - black-forest-labs/FLUX.1-schnellOpenAI - dall-e-3

from distilabel.pipeline import Pipeline
from distilabel.steps import KeepColumns
from distilabel.models.image_generation import InferenceEndpointsImageGeneration
from distilabel.steps.tasks import ImageGeneration

from datasets import load_dataset

ds = load_dataset("dvilasuero/finepersonas-v0.1-tiny", split="train").select(range(3))

with Pipeline(name="image_generation_pipeline") as pipeline:
    ilm = InferenceEndpointsImageGeneration(
        model_id="black-forest-labs/FLUX.1-schnell"
    )

    img_generation = ImageGeneration(
        name="flux_schnell",
        llm=ilm,
        input_mappings={"prompt": "persona"}
    )

    keep_columns = KeepColumns(columns=["persona", "model_name", "image"])

    img_generation >> keep_columns

提示的示例图像

一位对 19 世纪美国艺术和辛辛那提当地文化遗产感兴趣的当地艺术史学家和博物馆专业人士。

from distilabel.pipeline import Pipeline
from distilabel.steps import KeepColumns
from distilabel.models.image_generation import OpenAIImageGeneration
from distilabel.steps.tasks import ImageGeneration

from datasets import load_dataset

ds = load_dataset("dvilasuero/finepersonas-v0.1-tiny", split="train").select(range(3))

with Pipeline(name="image_generation_pipeline") as pipeline:
    ilm = OpenAIImageGeneration(
        model="dall-e-3",
        generation_kwargs={
            "size": "1024x1024",
            "quality": "standard",
            "style": "natural"
        }
    )

    img_generation = ImageGeneration(
        name="dalle-3"
        llm=ilm,
        input_mappings={"prompt": "persona"}
    )

    keep_columns = KeepColumns(columns=["persona", "model_name", "image"])

    img_generation >> keep_columns

提示的示例图像

一位对 19 世纪美国艺术和辛辛那提当地文化遗产感兴趣的当地艺术史学家和博物馆专业人士。

将 Distiset 保存为图像数据集

注意调用 Distiset.transform_columns_to_image，以便将图像直接上传为 Image dataset

if __name__ == "__main__":
    distiset = pipeline.run(use_cache=False, dataset=ds)
    # Save the images as `PIL.Image.Image`
    distiset = distiset.transform_columns_to_image("image")
    distiset.push_to_hub("plaguss/test-finepersonas-v0.1-tiny-flux-schnell")

完整的 pipeline 可以在以下示例中运行。请记住，您需要先安装 pillow：pip install distilabel[vision]。

运行

python examples/image_generation.py

image_generation.py

# Copyright 2023-present, Argilla, Inc.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     https://apache.ac.cn/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

from datasets import load_dataset

from distilabel.models.image_generation import InferenceEndpointsImageGeneration
from distilabel.pipeline import Pipeline
from distilabel.steps import KeepColumns
from distilabel.steps.tasks import ImageGeneration

ds = load_dataset("dvilasuero/finepersonas-v0.1-tiny", split="train").select(range(3))

with Pipeline(name="image_generation_pipeline") as pipeline:
    igm = InferenceEndpointsImageGeneration(model_id="black-forest-labs/FLUX.1-schnell")

    img_generation = ImageGeneration(
        name="flux_schnell",
        image_generation_model=igm,
        input_mappings={"prompt": "persona"},
    )

    keep_columns = KeepColumns(columns=["persona", "model_name", "image"])

    img_generation >> keep_columns


if __name__ == "__main__":
    distiset = pipeline.run(use_cache=False, dataset=ds)
    # Save the images as `PIL.Image.Image`
    distiset = distiset.transform_columns_to_image("image")
    distiset.push_to_hub("plaguss/test-finepersonas-v0.1-tiny-flux-schnell")

使用 distilabel 进行图像生成¶

使用 `distilabel` 进行图像生成¶