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
Upload sequential frames (min 10)
Drop File Here
- or -
Click to Upload
Choose Model
ConvLSTM
β-VAE
Prediction Mode
Single Frame
Multiple Frames
Number of Consecutive Frames (if Multiple Mode)
↺
1
20
Clear
Submit
Predicted Frames
Share via Link