layout生成之模型部署

1.现在选用的模型如下 多模态输入支持一张图片 从网络上下载图片
from http import HTTPStatus
import dashscope
from dashscope import Generation
from dashscope.api_entities.dashscope_response import Role

dashscope.api_key =""

def simple_multimodal_conversation_call(img,question):
    messages = [
        {
            "role": "user",
            "content": [
                {"image": f"{img}"},
                {"text": f"{question}"}
            ]
        }
    ]
    response = dashscope.MultiModalConversation.call(model='qwen-vl-plus',
                                                     messages=messages)
    if response.status_code == HTTPStatus.OK:
        print(response.output.choices[0]['message']['content'][0]['text'])
    else:
        print(response.code)
        print(response.message)

img = ''

text= ''

之前准备用chatgpt的接口 但花钱 先用通义千问测试

2.有想过先用图片转文字的模型+再加prompt发送 但后来发现有支持图片输入的模型

这个是图片转文字

from http import HTTPStatus
import dashscope


dashscope.api_key ="sk-a809926e939a412992ed4e36a1e1be12"
def call_with_messages():
    messages = [{'role': 'system', 'content': 'You are a helpful poster designer.'},
                {'role': 'user', 'content': '");'}]

    response = dashscope.Generation.call(
        dashscope.Generation.Models.qwen_turbo,
        messages=messages,
        result_format='message',  # set the result to be "message" format.
    )
    if response.status_code == HTTPStatus.OK:
        print(response)
    else:
        print('Request id: %s, Status code: %s, error code: %s, error message: %s' % (
            response.request_id, response.status_code,
            response.code, response.message
        ))

if __name__ == '__main__':
    call_with_messages()

3.识别多个图像 已弃用

img_captioning = pipeline(Tasks.image_captioning, model='damo/ofa_image-caption_coco_large_en', model_revision='v1.0.1')
result = img_captioning('https://shuangqing-public.oss-cn-zhangjiakou.aliyuncs.com/donuts.jpg')
print(result[OutputKeys.CAPTION]) # 'a bunch of donuts on a wooden board with popsicle sticks'
# 目前caption支持了batch inference：
result = img_captioning([{'image': 'https://shuangqing-public.oss-cn-zhangjiakou.aliyuncs.com/donuts.jpg'} for _ in range(3)], batch_size=2)
for r in result:
    print(r[OutputKeys.CAPTION])

4.一个字一个字输出

def sample_sync_call_streaming():
    # 设置需要生成的指令
    prompt_text = '用萝卜、土豆、茄子做饭，给我个菜谱。'
    # 调用dashscope.Generation.call方法生成响应流
    response_generator = dashscope.Generation.call(
        model='qwen-turbo',
        prompt=prompt_text,
        stream=True,
        top_p=0.8)

    head_idx = 0
    # 遍历响应流
    for resp in response_generator:
        # 获取每个响应中的文本段落
        paragraph = resp.output['text']
        # 打印文本段落中对应的文本
        print("\r%s" % paragraph[head_idx:len(paragraph)], end='')
        # 如果文本段落中存在换行符，则更新head_idx的值
        if (paragraph.rfind('\n') != -1):
            head_idx = paragraph.rfind('\n') + 1


# 调用sample_sync_call_streaming函数
sample_sync_call_streaming()