Image Details
Caption: Figure 3.
The architecture of the proposed model. The shapes in the figure, such as (32, 200, 1024), represent (batch_size, seq_len, nembd). Here, the input light-curve sequence length is fixed at seq_len = 200, the embedding dimension for the large-scale model is nembd = 1024, and the batch size is 32. The N in the figure denotes the number of stacked transformer blocks.
© 2025. The Author(s). Published by the American Astronomical Society.