# coding=utf-8
# Copyright 2022 EleutherAI and the HuggingFace Inc. team. All rights reserved.
#
# This code is based on EleutherAI's GPT-NeoX library and the GPT-NeoX
# and OPT implementations in this library. It has been modified from its
# original forms to accommodate minor architectural differences compared
# to GPT-NeoX and OPT used by the Meta AI team that trained the model.
#
# 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.
"""Sumi model configuration."""

from huggingface_hub.dataclasses import strict

from transformers.configuration_utils import PreTrainedConfig
from transformers.modeling_rope_utils import RopeParameters
from transformers.utils.type_validators import interval


@strict
class SumiConfig(PreTrainedConfig):
    r"""
    This is the configuration class for [`SumiModel`].

    Configuration objects inherit from [`PreTrainedConfig`] and can be used to control the model outputs. Read the
    documentation from [`PreTrainedConfig`] for more information.

    Args:
        vocab_size (`int`, *optional*, defaults to 32000):
            Vocabulary size of the Sumi model.
        hidden_size (`int`, *optional*, defaults to 4096):
            Dimension of the hidden representations.
        intermediate_size (`int`, *optional*, defaults to 11008):
            Dimension of the MLP representations.
        num_hidden_layers (`int`, *optional*, defaults to 32):
            Number of hidden layers in the Transformer encoder.
        num_attention_heads (`int`, *optional*, defaults to 32):
            Number of attention heads for each attention layer in the Transformer encoder.
        num_key_value_heads (`int`, *optional*):
            This is the number of key_value heads that should be used to implement Grouped Query Attention. If it is
            not specified, will default to `num_attention_heads`.
        hidden_act (`str` or `function`, *optional*, defaults to `"silu"`):
            The non-linear activation function (function or string) in the encoder.
        max_position_embeddings (`int`, *optional*, defaults to 2048):
            The maximum sequence length that this model might ever be used with.
        initializer_range (`float`, *optional*, defaults to 0.02):
            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
        rms_norm_eps (`float`, *optional*, defaults to 1e-06):
            The epsilon used by the rms normalization layers.
        use_cache (`bool`, *optional*, defaults to `False`):
            Whether or not the model should return the last key/values attentions. Disabled by
            default because the bidirectional diffusion model recomputes all key/values at every
            denoising step, so an incremental cache is not meaningful for generation.
        pad_token_id (`int`, *optional*):
            Padding token id.
        bos_token_id (`int`, *optional*, defaults to 1):
            Beginning of stream token id.
        eos_token_id (`int`, *optional*, defaults to 2):
            End of stream token id.
        pretraining_tp (`int`, *optional*, defaults to 1):
            Experimental feature. Tensor parallelism rank used during pretraining.
        tie_word_embeddings (`bool`, *optional*, defaults to `False`):
            Whether to tie weight embeddings.
        rope_parameters (`RopeParameters` or `dict`, *optional*):
            Dictionary containing the RoPE configuration. Holds `rope_theta` and (optionally) a `rope_type` together
            with its scaling parameters. The legacy `rope_theta` / `rope_scaling` fields found in older checkpoints are
            automatically migrated into this field by [`PreTrainedConfig`].
        attention_bias (`bool`, *optional*, defaults to `False`):
            Whether to use a bias in the query, key, value and output projection layers during self-attention.
        attention_dropout (`float`, *optional*, defaults to 0.0):
            The dropout ratio for the attention probabilities.
        mlp_bias (`bool`, *optional*, defaults to `False`):
            Whether to use a bias in up_proj, down_proj and gate_proj layers in the MLP layers.
        head_dim (`int`, *optional*):
            The attention head dimension. If None, it will default to hidden_size // num_attention_heads.
        add_qkv_bias (`bool`, *optional*, defaults to `False`):
            Whether to use bias terms only in the query, key, and value projection layers.
        uniform_diffusion_beta_is (`float`, *optional*, defaults to 1.0):
            Weight of the Itakura-Saito reconstruction term in the uniform-diffusion (GIDD) training loss
            (`SumiForMaskGeneration.forward` with `labels`). Matches the pretraining value.
        uniform_diffusion_z_loss_strength (`float`, *optional*, defaults to 1e-5):
            Coefficient of the auxiliary `logsumexp(logits) ** 2` z-loss added to the uniform-diffusion
            training loss for logit stability. Matches the pretraining value; set to `0.0` to disable.

    ```python
    >>> from transformers import AutoConfig

    >>> configuration = AutoConfig.from_pretrained("tohoku-nlp/open-uniform-diffusion", trust_remote_code=True)

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

    model_type = "sumi"
    keys_to_ignore_at_inference = ["past_key_values"]
    # Default tensor parallel plan for base model `SumiModel`
    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 = 32000
    hidden_size: int = 4096
    intermediate_size: int = 11008
    num_hidden_layers: int = 32
    num_attention_heads: int = 32
    num_key_value_heads: int | None = None
    hidden_act: str = "silu"
    max_position_embeddings: int = 2048
    initializer_range: float = interval(min=0.0, max=1.0)(default=0.02)
    rms_norm_eps: float = 1e-6
    use_cache: bool = False
    pad_token_id: int | None = None
    bos_token_id: int | None = 1
    eos_token_id: int | list[int] | None = 2
    pretraining_tp: int | None = 1
    tie_word_embeddings: bool = False
    rope_parameters: RopeParameters | dict | None = None
    attention_bias: bool = False
    attention_dropout: int | float | None = 0.0
    mlp_bias: bool = False
    head_dim: int | None = None
    add_qkv_bias: bool = False
    uniform_diffusion_beta_is: float = 1.0
    uniform_diffusion_z_loss_strength: float = 1e-5

    def __post_init__(self, **kwargs):
        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__ = ["SumiConfig"]