src.models.auto_clm.get_auto_clm_tokenizer

get_auto_clm_tokenizer(model_id: str, paths: Dict[str, Path], model_configs: Optional[dict] = None, use_pretrained_tokenizer: bool = True, reorder_and_upcast_attn: bool = True, scale_attn_by_inverse_layer_idx: bool = True, initial_weights: Optional[str] = None) → Tuple[AutoModelForCausalLM, PreTrainedTokenizer]

Download/Load AutoConfig and Instantiate Corresponding Model and Tokenizer.