React中加载TensorFlow.js模型时遭遇InputLayer配置与模型文件损坏错误的解决求助

我最近在React项目里尝试加载TensorFlow.js模型，写的代码是这样的：

const modelPromise = tf.loadLayersModel('/assets/models/tfjs_model/model.json')

可是每次尝试使用这个模型的时候，都会抛出下面的错误：

Model.jsx:10 Model load error _ValueError: An InputLayer should be passed either a `batchInputShape` or an `inputShape`.
    at new InputLayer (@tensorflow_tfjs.js?v=d977a120:19467:15)
    at fromConfig (@tensorflow_tfjs.js?v=d977a120:11535:12)
    at deserializeKerasObject (@tensorflow_tfjs.js?v=d977a120:17336:25)
    at deserialize (@tensorflow_tfjs.js?v=d977a120:20621:10)
    at processLayer (@tensorflow_tfjs.js?v=d977a120:22010:21)
    at fromConfig (@tensorflow_tfjs.js?v=d977a120:22024:7)
    at deserializeKerasObject (@tensorflow_tfjs.js?v=d977a120:17336:25)
    at deserialize (@tensorflow_tfjs.js?v=d977a120:20621:10)
    at loadLayersModelFromIOHandler (@tensorflow_tfjs.js?v=d977a120:24066:18)

我发现模型的model.json文件里有个batch_shape变量，于是试着把它改成batchInputShape，结果又碰到了新的报错：

Model.jsx:10 Model load error _ValueError: Corrupted configuration, expected array for nodeData: [object Object]
    at @tensorflow_tfjs.js?v=d977a120:22016:17
    at Array.forEach (<anonymous>)
    at processLayer (@tensorflow_tfjs.js?v=d977a120:22014:24)
    at fromConfig (@tensorflow_tfjs.js?v=d977a120:22024:7)
    at deserializeKerasObject (@tensorflow_tfjs.js?v=d977a120:17336:25)
    at deserialize (@tensorflow_tfjs.js?v=d977a120:20621:10)
    at loadLayersModelFromIOHandler (@tensorflow_tfjs.js?v=d977a120:24066:18)

现在我完全摸不着头绪，不知道该怎么解决这两个问题，有没有懂行的朋友能给我点建议？

内容来源于stack exchange