未处理的拒绝（错误）：model.execute(dict) 中提供的 dict['ImageTensor'] 的 dtype 必须为 int32，但为 float32

Question

我尝试运行 deep lab 模型 javascript 在视频上，但我收到错误未处理的拒绝（错误）：提供的 dict['ImageTensor'] 的数据类型在 model.execute(dict) 中必须是 int32，但是是 float32，这是我的代码

import React,{useEffect,useRef} from "react";
import Webcam from "react-webcam";

import * as tf from "@tensorflow/tfjs"
import * as deeplab from "@tensorflow-models/deeplab"


 
const Webcamera = () => {
    const webcamref = useRef(null)

    const loadModel = async () => {
        const modelName = 'ade20k';   // set to your preferred model, either `pascal`, `cityscapes` or `ade20k`
        const quantizationBytes = 2;  // either 1, 2 or 4
        const model = await deeplab.load({base: modelName, quantizationBytes});

        

        setInterval(() => {
            detect(model)
        },100)
      };

      const detect = async (model) => {
          if(typeof webcamref.current!=undefined && webcamref.current!=null && webcamref.current.video.readyState === 4 ){
              const video = webcamref.current.video
              const videowidth = webcamref.current.video.videowidth
              const videoheight = webcamref.current.video.videoheight

              webcamref.current.video.videowidth = videowidth
              webcamref.current.video.videoheight = videoheight
              

              model.segment(video)

              
          }
      }

      useEffect(()=>{loadModel()}, []);
      

return <Webcam ref={webcamref}/>
}

export default Webcamera

Answer 1

问题出在tensorflow版本上，请执行以下操作：卸载当前版本

npm uninstall @tensorflow/tfjs

安装低版本1.3.1

npm install @tensorflow/tfjs@1.3.1

Deeplab 似乎无法在更高版本上运行

未处理的拒绝（错误）：model.execute(dict) 中提供的 dict['ImageTensor'] 的 dtype 必须为 int32，但为 float32

Unhandled Rejection (Error): The dtype of dict['ImageTensor'] provided in model.execute(dict) must be int32, but was float32

machine-learning

reactjs

tensorflow.js