#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
#           This file was automatically generated from src/transformers/models/cwm/modular_cwm.py.
#               Do NOT edit this file manually as any edits will be overwritten by the generation of
#             the file from the modular. If any change should be done, please apply the change to the
#                          modular_cwm.py file directly. One of our CI enforces this.
#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
# Copyright 2025
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.


from huggingface_hub.dataclasses import strict

from ...configuration_utils import PreTrainedConfig
from ...utils import auto_docstring


@auto_docstring(checkpoint="facebook/cwm")
@strict
class CwmConfig(PreTrainedConfig):
    r"""
    ```python
    >>> from transformers import CwmModel, CwmConfig

    >>> # Initializing a Cwm cwm-7b style configuration
    >>> configuration = CwmConfig()

    >>> # Initializing a model from the cwm-7b style configuration
    >>> model = CwmModel(configuration)

    >>> # Accessing the model configuration
    >>> configuration = model.config
    ```"""

    model_type = "cwm"
    keys_to_ignore_at_inference = ["past_key_values"]
    # Default tensor parallel plan for base model `CwmModel`
    base_model_tp_plan = {
        "layers.*.self_attn.q_proj": "colwise",
        "layers.*.self_attn.k_proj": "colwise",
        "layers.*.self_attn.v_proj": "colwise",
        "layers.*.self_attn.o_proj": "rowwise",
        "layers.*.mlp.gate_proj": "colwise",
        "layers.*.mlp.up_proj": "colwise",
        "layers.*.mlp.down_proj": "rowwise",
    }
    base_model_pp_plan = {
        "embed_tokens": (["input_ids"], ["inputs_embeds"]),
        "layers": (["hidden_states", "attention_mask"], ["hidden_states"]),
        "norm": (["hidden_states"], ["hidden_states"]),
    }

    vocab_size: int = 128256
    hidden_size: int = 6144
    intermediate_size: int = 21504
    num_hidden_layers: int = 64
    num_attention_heads: int = 48
    num_key_value_heads: int = 8
    hidden_act: str = "silu"
    max_position_embeddings: int = 131072
    initializer_range: float = 0.02
    rms_norm_eps: float = 1e-5
    use_cache: bool = True
    pad_token_id: int | None = None
    bos_token_id: int = 128000
    eos_token_id: int | list[int] | None = None
    pretraining_tp: int = 1
    tie_word_embeddings: bool = False
    rope_parameters: dict | None = None
    attention_dropout: float | int = 0.0
    mlp_bias: bool = False
    head_dim: int = 128
    default_theta = 1_000_000.0
    sliding_window: int = 8192
    layer_types: list[str] | None = None  # ["full_attention"|"sliding_attention"] per layer

    def __post_init__(self, **kwargs):
        if self.rope_parameters is None:
            self.rope_parameters = {
                "rope_theta": 1_000_000.0,
                "factor": 16.0,
                "high_freq_factor": 4.0,
                "low_freq_factor": 1.0,
                "original_max_position_embeddings": 8192,
                "rope_type": "llama3",
            }

        if self.layer_types is None:
            # Default pattern: every 4th layer uses full attention, others use sliding attention
            window_pattern = 4
            self.layer_types = [
                ("full_attention" if (i % window_pattern == 0) else "sliding_attention")
                for i in range(self.num_hidden_layers)
            ]

        self.sliding_window = int(self.sliding_window) if self.sliding_window else None
        self.layer_types = list(self.layer_types)
        self.eos_token_id = self.eos_token_id if self.eos_token_id is not None else [128001, 128008, 128009]
        if self.head_dim is None:
            self.head_dim = self.hidden_size // self.num_attention_heads
        if self.num_key_value_heads is None:
            self.num_key_value_heads = self.num_attention_heads

        super().__post_init__(**kwargs)

    def validate_architecture(self):
        """Part of `@strict`-powered validation. Validates the architecture of the config."""
        if self.hidden_size % self.num_attention_heads != 0:
            raise ValueError(
                f"The hidden size ({self.hidden_size}) is not a multiple of the number of attention "
                f"heads ({self.num_attention_heads})."
            )


__all__ = ["CwmConfig"]