Instructions to use physical-intelligence/fast with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use physical-intelligence/fast with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("physical-intelligence/fast", dtype="auto") - Notebooks
- Google Colab
- Kaggle
Add pipeline tag, link to paper
Browse filesThis PR adds the pipeline tag `feature-extraction` so that the model can be found by the relevant filter on https://huggingface.co/models. It also
adds a link to the paper.
README.md
CHANGED
|
@@ -4,11 +4,12 @@ license: apache-2.0
|
|
| 4 |
tags:
|
| 5 |
- robotics
|
| 6 |
- tokenizer
|
|
|
|
| 7 |
---
|
| 8 |
|
| 9 |
# FAST: Efficient Action Tokenization for Vision-Language-Action Models
|
| 10 |
|
| 11 |
-
This is the official repo for the [FAST action tokenizer](https://www.pi.website/research/fast).
|
| 12 |
|
| 13 |
The action tokenizer maps any sequence of robot actions into a sequence of dense, discrete **action tokens** for training autoregressive VLA models.
|
| 14 |
|
|
|
|
| 4 |
tags:
|
| 5 |
- robotics
|
| 6 |
- tokenizer
|
| 7 |
+
pipeline_tag: feature-extraction
|
| 8 |
---
|
| 9 |
|
| 10 |
# FAST: Efficient Action Tokenization for Vision-Language-Action Models
|
| 11 |
|
| 12 |
+
This is the official repo for the [FAST action tokenizer](https://www.pi.website/research/fast) from the paper [FAST: Efficient Action Tokenization for Vision-Language-Action Models](https://huggingface.co/papers/2501.09747).
|
| 13 |
|
| 14 |
The action tokenizer maps any sequence of robot actions into a sequence of dense, discrete **action tokens** for training autoregressive VLA models.
|
| 15 |
|