许可证：apache-2.0
数据集：

• atasoglu/flickr8k-turkish
  语言：
• tr（土耳其语）
  评估指标：
• rouge
  库名称：transformers
  流水线标签：image-to-text
  标签：
• 图像转文本
• 图像描述生成
  基础模型：
• google/vit-base-patch16-224
• ytu-ce-cosmos/turkish-gpt2

vit-base-patch16-224-turkish-gpt2

该视觉编码器-解码器模型采用google/vit-base-patch16-224作为编码器，ytu-ce-cosmos/turkish-gpt2作为解码器，并在flickr8k-turkish数据集上进行了微调，用于生成土耳其语的图像描述。

使用方法

import torch  
from transformers import VisionEncoderDecoderModel, ViTImageProcessor, AutoTokenizer  
from PIL import Image  

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")  
model_id = "atasoglu/vit-base-patch16-224-turkish-gpt2"  
img = Image.open("example.jpg")  

feature_extractor = ViTImageProcessor.from_pretrained(model_id)  
tokenizer = AutoTokenizer.from_pretrained(model_id)  
model = VisionEncoderDecoderModel.from_pretrained(model_id)  
model.to(device)  

features = feature_extractor(images=[img], return_tensors="pt")  
pixel_values = features.pixel_values.to(device)  

generated_captions = tokenizer.batch_decode(  
    model.generate(pixel_values, max_new_tokens=20),  
    skip_special_tokens=True,  
)  

print(generated_captions)