YuhoLiang/facesyntheticsspigacaptioned

---
dataset_info:
features:
- name: image
dtype: image
- name: image_seg
dtype: image
- name: landmarks
dtype: string
- name: spiga
sequence:
sequence: float64
- name: spiga_seg
dtype: image
- name: image_caption
dtype: string
splits:
- name: train
num_bytes: 31087489990.0
num_examples: 100000
download_size: 31011261945
dataset_size: 31087489990.0
---
# Dataset Card for "face_synthetics_spiga_captioned"

This is a copy of the [Microsoft FaceSynthetics dataset with SPIGA-calculated landmark annotations](https://huggingface.co/datasets/pcuenq/face_synthetics_spiga), and additional BLIP-generated captions.
For a copy of the original FaceSynthetics dataset with no extra annotations, please refer to [pcuenq/face_synthetics](https://huggingface.co/datasets/pcuenq/face_synthetics).

Here is the code for parsing the dataset and generating the BLIP captions:
```py
from transformers import pipeline

dataset_name = "pcuenq/face_synthetics_spiga"
faces = load_dataset(dataset_name)
faces = faces["train"]

captioner = pipeline("image-to-text",model="Salesforce/blip-image-captioning-large", device=0)

def caption_image_data(example):
image = example["image"]
image_caption = captioner(image)[0]['generated_text']
example['image_caption'] = image_caption
return example

faces_proc = faces.map(caption_image_data)
faces_proc.push_to_hub(f"multimodalart/face_synthetics_spiga_captioned")
```

---
dataset_info: 数据集信息
features:
- name: image
dtype: 图像
- name: image_seg
dtype: 图像分割掩码
- name: landmarks
dtype: 字符串
- name: spiga
sequence:
sequence: float64
- name: spiga_seg
dtype: 图像分割掩码
- name: image_caption
dtype: 字符串
splits:
- name: train
num_bytes: 31087489990.0
num_examples: 100000
download_size: 31011261945
dataset_size: 31087489990.0
---
# 「face_synthetics_spiga_captioned」数据集卡片
本数据集为[Microsoft FaceSynthetics数据集（附带SPIGA计算得到的面部关键点（landmarks）标注）](https://huggingface.co/datasets/pcuenq/face_synthetics_spiga)的复刻版本，额外添加了由BLIP生成的图像标题。如需获取无额外标注的原始FaceSynthetics数据集，请参阅[pcuenq/face_synthetics](https://huggingface.co/datasets/pcuenq/face_synthetics)。

以下为解析该数据集并生成BLIP图像标题的代码示例：
py
from transformers import pipeline

dataset_name = "pcuenq/face_synthetics_spiga"
faces = load_dataset(dataset_name)
faces = faces["train"]

captioner = pipeline("image-to-text",model="Salesforce/blip-image-captioning-large", device=0)

def caption_image_data(example):
image = example["image"]
image_caption = captioner(image)[0]['generated_text']
example['image_caption'] = image_caption
return example

faces_proc = faces.map(caption_image_data)
faces_proc.push_to_hub(f"multimodalart/face_synthetics_spiga_captioned")