Instructions to use weikaih/molmo2-codec-pruning16-step1000 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use weikaih/molmo2-codec-pruning16-step1000 with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("weikaih/molmo2-codec-pruning16-step1000", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download config/native_p_tokenization.json from weikaih/molmo2-codec-pruning16-step1000: direct link, hf CLI and curl.
- Browser
- Download file 356 Bytes
-
https://huggingface.co/weikaih/molmo2-codec-pruning16-step1000/resolve/main/config/native_p_tokenization.json
- Command line
-
hf download hf://weikaih/molmo2-codec-pruning16-step1000/config/native_p_tokenization.json
-
curl -L -o native_p_tokenization.json https://huggingface.co/weikaih/molmo2-codec-pruning16-step1000/resolve/main/config/native_p_tokenization.json
356 Bytes
| { | |
| "connector": "independent-native-copy", | |
| "i_tokens_per_frame": 81, | |
| "mode": "pruning16", | |
| "p_input": "selected-rgb-from-motion-only-cache", | |
| "p_tokens_per_frame": 16, | |
| "rgb_persisted_in_cache": false, | |
| "schema": "molmo2-native-p-tokenization-v1", | |
| "selection": "robust-motion-residual-native-cell-top16", | |
| "vit": "shared-frozen-native-molmo" | |
| } | |