Instructions to use allenai/MolmoAct2-Think with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use allenai/MolmoAct2-Think with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoModelForImageTextToText model = AutoModelForImageTextToText.from_pretrained("allenai/MolmoAct2-Think", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
Commit History
-
Update paper link text
Verified signature94822d9 -
Update MolmoAct2 model card links and terminology
Verified signaturee28b96b
-
Update MolmoAct2 action mode inference API
Verified signaturef67c951
-
Remove unused action expert modules
Verified signature3d01bcf -
Fix MolmoAct2 remote norm stats loading
Verified signatureAllow predict_action to load norm_stats.json from the Hub when a checkpoint is loaded by repo id instead of a local directory.6490026 -
Update MolmoAct2 config schema and bf16 inference fix
Verified signatureRemove obsolete action expert config fields, keep max action dimensions/horizons explicit, and make bf16/fp16 action postprocessing safe during inference.fc8676c
-
Enable CUDA graph AR decode for discrete actions
Verified signatureGeneralizes the existing single-token decode CUDA graph helper and uses it for no-depth discrete action autoregressive decoding when enable_cuda_graph=True. The original KV-cache path is preserved when enable_cuda_graph=False.d82ea58
-
Add safety and citation sections
Verified signatureAppend model and hardware safety guidance plus the MolmoAct2 BibTeX citation to the model card.1bd5430 -
Add files using upload-large-folder tool
Verified signature048fc26
-
initial commit
Verified signature2478201