You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

B200 memory architecture checkpoints

This repository contains checkpoints from the streaming_3d_perception_online_memory project.

Architecture	Training step	Checkpoint	Configuration
GDN-2	20,000	`gdn2/checkpoint_000020000.pt`	`gdn2/config.yaml`
LaCT	20,000	`lact/checkpoint_000020000.pt`	`lact/config.yaml`
Mamba-3	20,000	`mamba3/checkpoint_000020000.pt`	`mamba3/config.yaml`
Transformer	13,000	`transformer/checkpoint_000013000.pt`	`transformer/config.yaml`

The runs used four B200 nodes with eight GPUs per node, for 32 B200 GPUs in total. These are complete training checkpoints and may contain model, optimizer, scheduler, and random-number-generator state.

See SHA256SUMS for checkpoint integrity hashes.

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