C# danbooru Stable Diffusion 提示词反推 Onnx Demo

本文链接：https://blog.csdn.net/weixin_46771779/article/details/136700826

本文介绍了如何使用C#和Onnx在Danbooru数据集上进行图像识别的示例，展示了从OpenCV读取图像、预处理、通过Onnx模型进行推理以及输出结果的过程。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

C# danbooru Stable Diffusion 提示词反推 Onnx Demo

说明

模型下载地址：https://huggingface.co/deepghs/ml-danbooru-onnx

效果

模型信息

Model Properties
-------------------------
---------------------------------------------------------------

Inputs
-------------------------
name：input
tensor：Float[-1, 3, -1, -1]
---------------------------------------------------------------

Outputs
-------------------------
name：output
tensor：Float[-1, 12547]
---------------------------------------------------------------

项目

代码

using Microsoft.ML.OnnxRuntime;
using Microsoft.ML.OnnxRuntime.Tensors;
using OpenCvSharp;
using System;
using System.Collections.Generic;
using System.Drawing;
using System.IO;
using System.Linq;
using System.Text;
using System.Windows.Forms;

namespace Onnx_Demo
{
    public partial class Form1 : Form
    {
        public Form1()
        {
            InitializeComponent();
        }

        string fileFilter = "*.*|*.bmp;*.jpg;*.jpeg;*.tiff;*.tiff;*.png";
        string image_path = "";
        DateTime dt1 = DateTime.Now;
        DateTime dt2 = DateTime.Now;
        string model_path;
        Mat image;

        SessionOptions options;
        InferenceSession onnx_session;
        Tensor<float> input_tensor;
        List<NamedOnnxValue> input_container;
        IDisposableReadOnlyCollection<DisposableNamedOnnxValue> result_infer;
        DisposableNamedOnnxValue[] results_onnxvalue;

        Tensor<float> result_tensors;

        StringBuilder sb = new StringBuilder();

        public string[] class_names;

        private void button1_Click(object sender, EventArgs e)
        {
            OpenFileDialog ofd = new OpenFileDialog();
            ofd.Filter = fileFilter;
            if (ofd.ShowDialog() != DialogResult.OK) return;
            pictureBox1.Image = null;
            image_path = ofd.FileName;
            pictureBox1.Image = new Bitmap(image_path);
            textBox1.Text = "";
            image = new Mat(image_path);
        }

        private void button2_Click(object sender, EventArgs e)
        {
            if (image_path == "")
            {
                return;
            }

            button2.Enabled = false;
            textBox1.Text = "";
            sb.Clear();
            Application.DoEvents();

            image = new Mat(image_path);

            // 将图片转为RGB通道
            Cv2.CvtColor(image, image, ColorConversionCodes.BGR2RGB);

            // 输入Tensor
            input_tensor = new DenseTensor<float>(new[] { 1, 3, image.Height, image.Width });

            // 输入Tensor
            for (int y = 0; y < image.Height; y++)
            {
                for (int x = 0; x < image.Width; x++)
                {
                    input_tensor[0, 0, y, x] = image.At<Vec3b>(y, x)[0] / 255f;
                    input_tensor[0, 1, y, x] = image.At<Vec3b>(y, x)[1] / 255f;
                    input_tensor[0, 2, y, x] = image.At<Vec3b>(y, x)[2] / 255f;
                }
            }

            //将 input_tensor 放入一个输入参数的容器，并指定名称
            input_container.Add(NamedOnnxValue.CreateFromTensor("input", input_tensor));

            dt1 = DateTime.Now;
            //运行 Inference 并获取结果
            result_infer = onnx_session.Run(input_container);
            dt2 = DateTime.Now;

            // 将输出结果转为DisposableNamedOnnxValue数组
            results_onnxvalue = result_infer.ToArray();

            // 读取第一个节点输出并转为Tensor数据
            result_tensors = results_onnxvalue[0].AsTensor<float>();

            var result_array = result_tensors.ToArray();

            double[] scores = new double[result_array.Length];
            for (int i = 0; i < result_array.Length; i++)
            {
                double score = 1 / (1 + Math.Exp(result_array[i] * -1));
                scores[i] = score;
            }

            List<ScoreIndex> ltResult = new List<ScoreIndex>();
            ScoreIndex temp;
            for (int i = 0; i < scores.Length; i++)
            {
                temp = new ScoreIndex(i, scores[i]);
                ltResult.Add(temp);
            }

            //根据分数倒序排序，取前10个
            var SortedByScore = ltResult.OrderByDescending(p => p.Score).ToList().Take(10);

            foreach (var item in SortedByScore)
            {
                sb.Append(class_names[item.Index] + ",");
            }
            sb.Length--; // 将长度减1来移除最后一个字符

            sb.AppendLine("");
            sb.AppendLine("------------------");

            // 只取分数最高的
            // float max = result_array.Max();
            // int maxIndex = Array.IndexOf(result_array, max);
            // sb.AppendLine(class_names[maxIndex]+" "+ max.ToString("P2"));

            sb.AppendLine("推理耗时:" + (dt2 - dt1).TotalMilliseconds + "ms");
            textBox1.Text = sb.ToString();
            button2.Enabled = true;
        }

        private void Form1_Load(object sender, EventArgs e)
        {
            model_path = "model/ml_danbooru.onnx";

            // 创建输出会话，用于输出模型读取信息
            options = new SessionOptions();
            options.LogSeverityLevel = OrtLoggingLevel.ORT_LOGGING_LEVEL_INFO;
            options.AppendExecutionProvider_CPU(0);// 设置为CPU上运行

            // 创建推理模型类，读取本地模型文件
            onnx_session = new InferenceSession(model_path, options);//model_path 为onnx模型文件的路径

            // 创建输入容器
            input_container = new List<NamedOnnxValue>();

            image_path = "test_img/2.jpg";
            pictureBox1.Image = new Bitmap(image_path);
            image = new Mat(image_path);

            List<string> str = new List<string>();
            StreamReader sr = new StreamReader("model/lable.txt");
            string line;
            while ((line = sr.ReadLine()) != null)
            {
                str.Add(line);
            }
            class_names = str.ToArray();
        }

    }
}