Models

Some experiments of tfts in Kaggle Dataset

Models supported

You can use below models with AutoModel

• RNN

• Seq2seq

• TCN

• WaveNet

• Bert

• Transformer

• DLinear

• NBeats

• AutoFormer

• Informer

config = AutoConfig.for_model("seq2seq")
model = AutoModel.from_config(config, predict_sequence_length=predict_sequence_length)

Add a custom head for tfts model

config = AutoConfig.for_model("seq2seq")
model = AutoModel.from_config(config, predict_sequence_length=predict_sequence_length)
model.project = tf.keras.Sequential(
    layers=[],
    trainable=True,
    name=None
)