Pong Frame Predictor (ConvLSTM / β-VAE)

🕹️ Pong Frame Prediction

Upload 10 sequential Pong frames and select a model + prediction mode.

  • ConvLSTM → Learns temporal dynamics directly in pixel space
  • β-VAE → Predicts next frames via latent-space reconstruction
Choose Model
Prediction Mode
1 20