_target_: text_recognizer.networks.vqvae.pixelcnn.Encoder
in_channels: 1 
hidden_dim: 8
channels_multipliers: [1, 2, 8]
dropout_rate: 0.25