Tactile RoPE DINO encoders

Two trained S-size tactile encoders using depth_deform_raw input.

Version	Encoder weights
V1	DINO_TactileRoPE/encoder.pth
V2	DINO_TactileRoPE_V2/encoder.pth

These are the selected encoder exports, not the full last.pth training checkpoints. V1 and V2 use different architectures and are not interchangeable.

Code and inference utilities: https://github.com/Jeffry-wen/UniVTAC-Positional_Encoding

File sizes, SHA-256 checksums and training run identifiers are recorded in weights_manifest.json.

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