Qwen2-VL-7B模型本地安装教程

1. 准备环境

确保您的Python版本为3.10，PyTorch版本为2.2.0，CUDA版本为12.1.1。

# 检查Python版本
python --version
​
# 检查PyTorch版本
python -c "import torch; print(torch.__version__)"
​
# 检查CUDA版本
nvcc --version

2. 安装transformers和accelerate

通过pip安装transformers和accelerate：

pip install git+https://github.com/huggingface/transformers accelerate

3. 安装工具集

通过pip安装qwen-vl-utils：

pip install qwen-vl-utils

4. 编写测试文件

创建一个Python文件（例如test_qwen2_vl_7b.py），并编写以下代码：

from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info
​
# default: Load the model on the available device(s)
model = Qwen2VLForConditionalGeneration.from_pretrained(
    "Qwen/Qwen2-VL-7B-Instruct", torch_dtype="auto", device_map="auto"
)
​
# We recommend enabling flash_attention_2 for better acceleration and memory saving, especially in multi-image and video scenarios.
# model = Qwen2VLForConditionalGeneration.from_pretrained(
#     "Qwen/Qwen2-VL-7B-Instruct",
#     torch_dtype=torch.bfloat16,
#     attn_implementation="flash_attention_2",
#     device_map="auto",
# )
​
# default processer
processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-7B-Instruct")
​
# The default range for the number of visual tokens per image in the model is 4-16384. You can set min_pixels and max_pixels according to your needs, such as a token count range of 256-1280, to balance speed and memory usage.
# min_pixels = 256*28*28
# max_pixels = 1280*28*28
# processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-7B-Instruct", min_pixels=min_pixels, max_pixels=max_pixels)
​
messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "image",
                "image": "https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg",
            },
            {"type": "text", "text": "Describe this image."},
        ],
    }
]
​
# 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: Generation of the output
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)
​

5. 执行测试文件

在终端中运行以下命令以执行测试文件：

python test_qwen2_vl_7b.py

注意点

• 确保您的显卡显存至少为40GB。使用24GB的3090显卡可能会导致显存不足的错误。