Instructions to use INSAIT-Institute/arvla-bridge with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use INSAIT-Institute/arvla-bridge with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("INSAIT-Institute/arvla-bridge", trust_remote_code=True, dtype="auto") - Notebooks
- Google Colab
- Kaggle
| { | |
| "processor_class": "VLARMHFProcessor", | |
| "base_processor_model_id": "google/paligemma-3b-mix-224", | |
| "vlarm_processor_config": { | |
| "control_io_config": { | |
| "future_controls_sequence_length": 5, | |
| "future_controls_sequence_stride_sec": 0.2, | |
| "future_frames_sequence_length": 1, | |
| "future_frames_sequence_stride_sec": null, | |
| "future_control_offset_sec": 0.0, | |
| "past_scalars_sequence_length": 1, | |
| "past_frames_sequence_length": 1, | |
| "past_scalars_stride_sec": null, | |
| "past_frames_stride_sec": null, | |
| "sequence_frames": 1, | |
| "sequence_frames_stride_sec": null | |
| }, | |
| "obs_translation_norm": "bounds_q99", | |
| "obs_rotation_norm": "none", | |
| "translation_norm": { | |
| "high": [ | |
| 0.04, | |
| 0.04, | |
| 0.04 | |
| ], | |
| "low": [ | |
| -0.04, | |
| -0.04, | |
| -0.04 | |
| ] | |
| }, | |
| "rotation_norm": "none", | |
| "joints_norm": { | |
| "high": [ | |
| 3.141592653589793, | |
| 3.141592653589793, | |
| 3.141592653589793, | |
| 3.141592653589793, | |
| 3.141592653589793, | |
| 3.141592653589793, | |
| 3.141592653589793 | |
| ], | |
| "low": [ | |
| -3.141592653589793, | |
| -3.141592653589793, | |
| -3.141592653589793, | |
| -3.141592653589793, | |
| -3.141592653589793, | |
| -3.141592653589793, | |
| -3.141592653589793 | |
| ] | |
| }, | |
| "rotation_format": "quaternion", | |
| "eef_control_frame": false, | |
| "delta_controls": true, | |
| "image_resize": "naive", | |
| "max_language_tokens": 75, | |
| "control_tokenizer_config": {}, | |
| "control_stats_path": "/home/yutong/experiments/vlams/vlarm/sess_2025_11_14_11_29_01_msp3-3_yutong_hu_fast_vl+pt_ar.6/hf_export/hare-determined-crimson-transformers/stats/control_stats.yaml", | |
| "observation_stats_path": "/home/yutong/experiments/vlams/vlarm/sess_2025_11_14_11_29_01_msp3-3_yutong_hu_fast_vl+pt_ar.6/hf_export/hare-determined-crimson-transformers/stats/observation_stats.yaml" | |
| }, | |
| "vlm_processor_config": { | |
| "image_token": "<image>", | |
| "image_sizes": { | |
| "main": { | |
| "width": 224, | |
| "height": 224 | |
| } | |
| } | |
| }, | |
| "auto_map": { | |
| "AutoProcessor": "processing_vlarm_hf.VLARMHFProcessor" | |
| } | |
| } |