参考：
https://huggingface.co/Qwen/Qwen2-VL-2B-Instruct

模型：

export HF_ENDPOINT=https://hf-mirror.com

huggingface-cli download --resume-download --local-dir-use-symlinks False Qwen/Qwen2-VL-2B-Instruct  --local-dir qwen2-vl

安装：
transformers-4.45.0.dev0
accelerate-0.34.2 safetensors-0.4.5

pip install git+https://github.com/huggingface/transformers
pip install 'accelerate>=0.26.0'

代码：

单张图片

from PIL import Image
import requests
import torch
from torchvision import io
from typing import Dict
from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor

# Load the model in half-precision on the available device(s)
model = Qwen2VLForConditionalGeneration.from_pretrained(
    "/ai/qwen2-vl", torch_dtype="auto", device_map="auto"
)
processor = AutoProcessor.from_pretrained("/ai/qwen2-vl")




# Image
url = "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg"
image = Image.open(requests.get(url, stream=True).raw)

conversation = [
    {
        "role": "user",
        "content": [
            {
                "type": "image",
            },
            {"type": "text", "text": "Describe this image."},
        ],
    }
]


# Preprocess the inputs
text_prompt = processor.apply_chat_template(conversation, add_generation_prompt=True)
# Excepted output: '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n<|im_start|>user\n<|vision_start|><|image_pad|><|vision_end|>Describe this image.<|im_end|>\n<|im_start|>assistant\n'

inputs = processor(
    text=[text_prompt], images=[image], padding=True, return_tensors="pt"
)
inputs = inputs.to("cuda")

# Inference: Generation of the output
output_ids = model.generate(**inputs, max_new_tokens=128)
generated_ids = [
    output_ids[len(input_ids) :]
    for input_ids, output_ids in zip(inputs.input_ids, output_ids)
]
output_text = processor.batch_decode(
    generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True
)
print(output_text)

这是图片：

中文问

# Image
url = "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg"
image = Image.open(requests.get(url, stream=True).raw)

conversation = [
    {
        "role": "user",
        "content": [
            {
                "type": "image",
            },
            {"type": "text", "text": "描述下这张图片."},
        ],
    }
]


# Preprocess the inputs
text_prompt = processor.apply_chat_template(conversation, add_generation_prompt=True)
# Excepted output: '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n<|im_start|>user\n<|vision_start|><|image_pad|><|vision_end|>Describe this image.<|im_end|>\n<|im_start|>assistant\n'

inputs = processor(
    text=[text_prompt], images=[image], padding=True, return_tensors="pt"
)
inputs = inputs.to("cuda")
# Inference: Generation of the output
output_ids = model.generate(**inputs, max_new_tokens=128)
generated_ids = [
    output_ids[len(input_ids) :]
    for input_ids, output_ids in zip(inputs.input_ids, output_ids)
]
output_text = processor.batch_decode(
    generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True
)
print(output_text)

多张图片

def load_images(image_info):
    images = []
    for info in image_info:
        if "image" in info:
            if info["image"].startswith("http"):
                image = Image.open(requests.get(info["image"], stream=True).raw)
            else:
                image = Image.open(info["image"])
            images.append(image)
    return images

# Messages containing multiple images and a text query
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "image": "/ai/fight.png"},
            {"type": "image", "image": "/ai/long.png"},
            {"type": "text", "text": "描述下这两张图片"},
        ],
    }
]

# Load images
image_info = messages[0]["content"][:2]  # Extract image info from the message
images = load_images(image_info)

# Preprocess the inputs
text_prompt = processor.apply_chat_template(messages, add_generation_prompt=True)

inputs = processor(
    text=[text_prompt], images=images, padding=True, return_tensors="pt"
)
inputs = inputs.to("cuda")

# Inference: Generation of the output
output_ids = model.generate(**inputs, max_new_tokens=128)
generated_ids = [
    output_ids[len(input_ids) :]
    for input_ids, output_ids in zip(inputs.input_ids, output_ids)
]
output_text = processor.batch_decode(
    generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True
)
print(output_text)

视频
安装

pip install qwen-vl-utils

from qwen_vl_utils import process_vision_info

# Messages containing a images list as a video and a text query
messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "video",
                "video": [
                    "file:///path/to/frame1.jpg",
                    "file:///path/to/frame2.jpg",
                    "file:///path/to/frame3.jpg",
                    "file:///path/to/frame4.jpg",
                ],
                "fps": 1.0,
            },
            {"type": "text", "text": "Describe this video."},
        ],
    }
]
# Messages containing a video and a text query
messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "video",
                "video": "/ai/血液从上肢流入上腔静脉.mp4",
                "max_pixels": 360 * 420,
                "fps": 1.0,
            },
            {"type": "text", "text": "描述下这个视频"},
        ],
    }
]

# Preparation for inference
text = processor.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True
)
image_inputs, video_inputs = process_vision_info(messages)
inputs = processor(
    text=[text],
    images=image_inputs,
    videos=video_inputs,
    padding=True,
    return_tensors="pt",
)
inputs = inputs.to("cuda")

# Inference
generated_ids = model.generate(**inputs, max_new_tokens=128)
generated_ids_trimmed = [
    out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
    generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text)