多模态情感分析

使用 Hugging Face 的多模态模型

James Chapman

Curriculum Manager, DataCamp

视觉-语言模型（VLM）

通用视觉-语言模型架构示意图

图像与文本分别编码
特征融合
生成共享表示
- 一模多用，覆盖多任务

视觉推理任务

VQA：图像 + 问题 → 直接回答图像内容

视觉问答示例：热狗

¹ https://arxiv.org/pdf/1505.00468

视觉推理任务

VQA：图像 + 问题 → 直接回答图像内容
匹配：图像 + 陈述 → 真/假判定

匹配示例：热狗与相关陈述

视觉推理任务

VQA：图像 + 问题 → 直接回答图像内容
匹配：图像 + 陈述 → 真/假判定
蕴含：图像 + 对比文本 → 语义关系检验

蕴含示例：热狗，含蕴含、中立、矛盾假设

用例：股价影响

from datasets import load_dataset
dset = "RealTimeData/bbc_news_alltime"
dataset = load_dataset(dset, '2017-01', 
                       split="train")


image = dataset[87]["top_image"]
content = dataset[87]["content"]
print(content)

'Ford\'s decision to cancel a $1.6bn 
investment in Mexico and 
invest an extra $700m in Michigan ...

新闻文章中的图片：汽车生产投资

Qwen 2 VLMs

from transformers import Qwen2VLForConditionalGeneration
from qwen_vl_utils import process_vision_info


vl_model = Qwen2VLForConditionalGeneration.from_pretrained(
    "Qwen/Qwen2-VL-2B-Instruct", device_map="auto", torch_dtype="auto"
)

预处理器

from transformers import Qwen2VLProcessor

min_pixels = 224 * 224
max_pixels = 448 * 448
vl_model_processor = Qwen2VLProcessor.from_pretrained(
    "Qwen/Qwen2-VL-2B-Instruct", min_pixels=min_pixels, max_pixels=max_pixels
)

多模态提示

text_query = f"Is the sentiment of the following content good or bad for the Ford share price:
{article_text}. Provide reasoning."


chat_template = [
    {
        "role": "user",
        "content": [
            {"type": "image", "image": article_image},
            {"type": "text", "text": text_query}
        ]
    }
]

VLM 分类

text = vl_model_processor.apply_chat_template(chat_template, tokenize=False,
                                              add_generation_prompt=True)

image_inputs, _ = process_vision_info(chat_template)


inputs = vl_model_processor(text=[text], images=image_inputs, padding=True,
                            return_tensors="pt")


generated_ids = vl_model.generate(**inputs, max_new_tokens=500)

generated_ids_trimmed = [out_ids[len(in_ids) :] for in_ids, out_ids\
                         in zip(inputs.input_ids, generated_ids)]

VLM 分类

output_text = vl_model_processor.batch_decode(
    generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text[0])

The sentiment of the provided text is negative. The author is expressing concern and
skepticism ...

Vamos praticar!

使用 Hugging Face 的多模态模型