unity sentis 2.1.0

unity sentis 2.1.0学习记录帖

unity sentis 2.1.0手写字体识别

---

---

前言

随着unity6的发布 记录自己学习历程

---

一、使用步骤

1. 引入库

代码如下（省去安装步骤）：

using UnityEngine;
using Unity.Sentis;

2. 将 softmax 层添加到模型的末尾

代码如下（示例）：

FunctionalGraph graph = new FunctionalGraph();
// Set the inputs of the graph from the original model inputs and return an array of functional tensors
FunctionalTensor[] inputs = graph.AddInputs(runtimeModel);
FunctionalTensor[] outputs = Functional.Forward(runtimeModel, inputs);
// Calculate the softmax of the first output with the functional API.
FunctionalTensor softmax = Functional.Softmax(outputs[0]);

// Create a model with softmax by compiling the functional graph.
runtimeModel = graph.Compile(softmax);

3. 实现

// 将输入数据创建为张量
Tensor inputTensor = TextureConverter.ToTensor(inputTexture, width: 28, height: 28, channels: 1);
// 创建引擎
worker = new Worker(runtimeModel, BackendType.GPUCompute);
// 使用输入数据运行模型
worker.Schedule(inputTensor);
//Get the result
Tensor<float> outputTensor = worker.PeekOutput() as Tensor<float>;
results = outputTensor.DownloadToArray();
Debug.Log("识别结果：" + GetValue(results));

private int GetValue(float[] results)
{
    int index = 0;
    float value = 0;
    for (int i = 0; i < results.Length; i++)
    {
        if (value < results[i])
        {
            value = results[i];
            index = i;
        }
    }
    return index;
}

调整输出得到结果，索引值为结果值

---

总结

just soso