Adaptive 3D-RoPE checkpoints

Six PyTorch checkpoints for the public Adaptive 3D-RoPE code repository. Each directory contains a checkpoint-149.pth and matches the profile with the same name in position_encoding_profiles.yaml.

The weights are trained on externally prepared CSI data. This repository does not contain raw or processed datasets. Download the code, install its pinned dependencies, prepare DATA_ROOT, and evaluate with python cli.py eval --profile PROFILE ....

Use the SHA-256 values in checksums.sha256 before loading a checkpoint. Do not mix profiles: the code performs strict state-dict and saved-argument compatibility checks.

Source commit: f0df7e6404e95d78f24824f3582d64f9967d5149.

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