DreamDojo Jokeru 2B Self-Forcing

Self-Forcing/DMD student continued from iter_000000100 for an approximately two-hour 8-GPU run.

Uploaded checkpoint: iter_000000700
Base post-trained model: DreamDojo Jokeru 2B iter_000003000
Sampler: 4-step Self-Forcing
Resolution: 480x640
Action dimension: 384
Training command exit status: 0
Post-training inference validation exit status: 0

The repository stores model DCP shards only; optimizer and scheduler states are intentionally omitted. Load from checkpoints/iter_000000700 and provide a local Wan2.1 VAE checkpoint for inference.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support