# coding=utf-8
# Copyright The HuggingFace Inc. team. All rights reserved.
#
# 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.
""" XGLM model configuration"""

from ...configuration_utils import PretrainedConfig
from ...utils import logging


logger = logging.get_logger(__name__)


from ..deprecated._archive_maps import XGLM_PRETRAINED_CONFIG_ARCHIVE_MAP  # noqa: F401, E402


class XGLMConfig(PretrainedConfig):
    r"""
    This is the configuration class to store the configuration of a [`XGLMModel`]. It is used to instantiate an XGLM
    model according to the specified arguments, defining the model architecture. Instantiating a configuration with the
    defaults will yield a similar configuration to that of the XGLM
    [facebook/xglm-564M](https://huggingface.co/facebook/xglm-564M) architecture.

    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 256008):
            Vocabulary size of the XGLM model. Defines the number of different tokens that can be represented by the
            `inputs_ids` passed when calling [`XGLMModel`] or [`FlaxXGLMModel`].
        max_position_embeddings (`int`, *optional*, defaults to 2048):
            The maximum sequence length that this model might ever be used with. Typically set this to something large
            just in case (e.g., 512 or 1024 or 2048).
        d_model (`int`, *optional*, defaults to 1024):
            Dimension of the layers and the pooler layer.
        ffn_dim (`int`, *optional*, defaults to 4096):
            Dimension of the "intermediate" (often named feed-forward) layer in decoder.
        num_layers (`int`, *optional*, defaults to 24):
            Number of hidden layers Transformer decoder.
        attention_heads (`int`, *optional*, defaults to 16):
            Number of attention heads for each attention layer in the Transformer decoder.
        activation_function (`str` or `function`, *optional*, defaults to `"gelu"`):
            The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`,
            `"relu"`, `"silu"` and `"gelu_new"` are supported.
        dropout (`float`, *optional*, defaults to 0.1):
            The dropout probability for all fully connected layers in the embeddings, dencoder, and pooler.
        attention_dropout (`float`, *optional*, defaults to 0.1):
            The dropout ratio for the attention probabilities.
        activation_dropout (`float`, *optional*, defaults to 0.0):
            The dropout ratio for activations inside the fully connected layer.
        layerdrop (`float`, *optional*, defaults to 0.0):
            The LayerDrop probability for the encoder. See the [LayerDrop paper](see https://arxiv.org/abs/1909.11556)
            for more details.
        init_std (`float`, *optional*, defaults to 0.02):
            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
        scale_embedding (`bool`, *optional*, defaults to `True`):
            Scale embeddings by diving by sqrt(d_model).
        use_cache (`bool`, *optional*, defaults to `True`):
            Whether or not the model should return the last key/values attentions (not used by all models).

    Example:

    ```python
    >>> from transformers import XGLMModel, XGLMConfig

    >>> # Initializing a XGLM facebook/xglm-564M style configuration
    >>> configuration = XGLMConfig()

    >>> # Initializing a model from the facebook/xglm-564M style configuration
    >>> model = XGLMModel(configuration)

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

    model_type = "xglm"
    keys_to_ignore_at_inference = ["past_key_values"]

    attribute_map = {
        "num_attention_heads": "attention_heads",
        "hidden_size": "d_model",
        "num_hidden_layers": "num_layers",
    }

    def __init__(
        self,
        vocab_size=256008,
        max_position_embeddings=2048,
        d_model=1024,
        ffn_dim=4096,
        num_layers=24,
        attention_heads=16,
        activation_function="gelu",
        dropout=0.1,
        attention_dropout=0.1,
        activation_dropout=0.0,
        layerdrop=0.0,
        init_std=0.02,
        scale_embedding=True,
        use_cache=True,
        decoder_start_token_id=2,
        pad_token_id=1,
        bos_token_id=0,
        eos_token_id=2,
        **kwargs,
    ):
        self.vocab_size = vocab_size
        self.max_position_embeddings = max_position_embeddings
        self.d_model = d_model
        self.ffn_dim = ffn_dim
        self.num_layers = num_layers
        self.attention_heads = attention_heads
        self.activation_function = activation_function
        self.dropout = dropout
        self.attention_dropout = attention_dropout
        self.activation_dropout = activation_dropout
        self.layerdrop = layerdrop
        self.init_std = init_std
        self.scale_embedding = scale_embedding  # scale factor will be sqrt(d_model) if True
        self.use_cache = use_cache

        super().__init__(
            pad_token_id=pad_token_id,
            bos_token_id=bos_token_id,
            eos_token_id=eos_token_id,
            decoder_start_token_id=decoder_start_token_id,
            **kwargs,
        )