Efficient-Large-Model/Fast-dDrive

1"""Fast-dDrive HF release package.2 3Mirrors the layout of ``Efficient-Large-Model/Fast_dVLM_3B`` on the Hugging Face4Hub: a single :mod:`modeling` module that holds the model definition plus5inference-time decoding paths (Section Diffusion, Scaffold Spec, and6Scaffold Spec with multi-trajectory rollouts), and a :mod:`configuration`7module with the config classes.8 9Users normally load the model via::10 11    from transformers import AutoModelForCausalLM, AutoTokenizer12    model = AutoModelForCausalLM.from_pretrained(13        "Efficient-Large-Model/Fast_dDrive",  # or local path14        trust_remote_code=True,15    )16 17with the ``auto_map`` entry in :file:`config.json` pointing back to the classes18defined here.19"""20 21from .configuration import (22    Fast_dDriveConfig,23    Fast_dDriveTextConfig,24    Fast_dDriveVisionConfig,25)26from .modeling import Fast_dDriveForConditionalGeneration27 28__all__ = [29    "Fast_dDriveConfig",30    "Fast_dDriveTextConfig",31    "Fast_dDriveVisionConfig",32    "Fast_dDriveForConditionalGeneration",33]34