[英]Keras function api, setting weight manually to a layer
In keras Sequential model, one can set weight directly using set_weights
method. 在keras顺序模型中,可以使用
set_weights
方法直接设置权重。
model.layers[n].set_weights([your_wight])
However I am facing problem if I am trying to set weight to a layer using functional API. 但是,如果我尝试使用功能性API为图层设置权重,则会遇到问题。
Here is the code snippet: 这是代码片段:
emb = Embedding(max_words, embedding_dim, input_length=maxlen)(merge_ip)
#skipping some lines
.
.
emb.set_weights([some_weight_matrix])
This is throwing error that 这引发了错误
AttributeError: 'Tensor' object has no attribute 'set_weights'
which I think becouse emb
is a Tensor object. 我认为因为
emb
是一个Tensor对象。
I am wondering how to set wight properly in my model 我想知道如何在模型中正确设置体重
If you want to set the weights on Embedding layers you might add them to the constructor like this: 如果要在“嵌入”层上设置权重,可以将其添加到构造函数中,如下所示:
from keras.layers import Embedding
embedding_layer = Embedding(len(word_index) + 1,
EMBEDDING_DIM,
weights=[embedding_matrix],
input_length=MAX_SEQUENCE_LENGTH,
trainable=False)
https://blog.keras.io/using-pre-trained-word-embeddings-in-a-keras-model.html https://blog.keras.io/using-pre-trained-word-embeddings-in-a-keras-model.html
Later then you can hand over merge_ip
: 之后,您可以移交
merge_ip
:
x = embedding_layer(merge_ip)
embed_layer = Embedding(max_words, embedding_dim, input_length=maxlen)
emp = embed_layer(merge_ip)
embed_layer.set_weights("...")
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.