在这段代码的基础上，将gru换成bigru，并在bigru后面加入多头注意力机制

好的，以下是修改后的代码： ```python from keras.layers import Input, Dense, Bidirectional, CuDNNGRU, SeqSelfAttention, MultiHeadAttention from keras.models import Model def create_model(input_shape, num_heads=8): input_layer = Input(shape=input_shape) x = Bidirectional(CuDNNGRU(64, return_sequences=True))(input_layer) x = SeqSelfAttention(attention_activation='sigmoid')(x) x, attn_weights = MultiHead