Instructions to use pacifio/untoken-v2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use pacifio/untoken-v2 with Transformers:
# Load model directly from transformers import UntokenCompressor model = UntokenCompressor.from_pretrained("pacifio/untoken-v2", dtype="auto") - Notebooks
- Google Colab
- Kaggle
| { | |
| "architectures": [ | |
| "UntokenCompressor" | |
| ], | |
| "dropout": 0.1, | |
| "dtype": "float32", | |
| "head_hidden": 256, | |
| "model_type": "untoken_compressor", | |
| "target_ratio": 0.3, | |
| "transformers_version": "5.2.0" | |
| } | |