我正在尝试解析json 文件以在 JuliaKeras
中创建模型( v1.1.0 和v0.10.4)。Flux
Keras
Flux
Dense
层配置示例:
{
"class_name": "Dense",
"config": {
"W_constraint": null,
"b_constraint": null,
"name": "dense_1",
"output_dim": 512,
"activity_regularizer": null,
"trainable": true,
"init": "glorot_normal",
"bias": true,
"input_dtype": "float32",
"input_dim": 4096,
"b_regularizer": null,
"W_regularizer": {
"l2": 0.0010000000474974513,
"name": "WeightRegularizer",
"l1": 0.0
},
"activation": "relu",
"batch_input_shape": [null, 4096]
}
}
因此,我很清楚如何在Flux
. 但是W_constraint
和W_regularizer
呢?Flux
我在'sDense
层没有发现任何类似的东西。它存在吗?我应该自己实施吗?层的这些参数是否Dense
甚至很重要,或者在创建模型时可以轻松跳过Flux
而不会严重改变性能?