Robotics
PyTorch
Cosmos
xperience10m_task_baseline_suite
embodied-ai
multimodal
xperience-10m
baseline
evaluation
qwen3-omni
Instructions to use cy0307/ropedia-xperience-10m-task-baselines with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Cosmos
How to use cy0307/ropedia-xperience-10m-task-baselines with Cosmos:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
File size: 698 Bytes
540e67a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 | # Qwen3-Omni Adapter Smoke Test
- Base model target: `Qwen/Qwen3-Omni-30B-A3B-Thinking`
- Qwen3-Omni weights loaded: `false`
- Episodes: `1`
- Windows: `59` total, `41` train, `18` test
- Split: `single_episode_chronological`
- Feature dimension: `4262`
- Adapter soft-token blocks: `11`
- Accuracy: `0.0000`
- Macro-F1: `0.0000`
## Why this is the minimum real test
This run uses real Ropedia annotation/video-derived feature blocks. It tests the sensor-adapter side that depth, pose, mocap, contacts, and IMU need before those tokens are attached to Qwen3-Omni. It deliberately avoids downloading the 30B Qwen3-Omni weights until the data path, labels, splits, and storage plan are confirmed.
|