configuration_dense_gau_retnet.py

# 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.
""" DenseGauRetNet model configuration"""

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

logger = logging.get_logger(__name__)

DenseGauRetNet_PRETRAINED_CONFIG_ARCHIVE_MAP = {}


class DenseGauRetNetConfig(PretrainedConfig):
    model_type = "DenseGauRetNet"
    _auto_class = "AutoConfig"

    def __init__(
            self,
            hidden_act: str = "silu",
            hidden_size: int = 1536,
            query_key_dim: int = 768,
            initializer_range: float = 0.02,
            max_position_embeddings: int = 2048,
            num_attention_heads: int = 2,
            num_hidden_layers: int = 16,
            rms_norm_eps: float = 1e-06,
            layernorm_eps: float = 1e-5,
            retnorm: bool = False,
            vocab_size: int = 32001,
            v_factor: int = 2,
            intermediate_k_select_scale: int = 8,
            intermediate_v_select_scale: int = 32,
            dense_block_layers: int = 2,
            dropout: float = 0.1,
            use_cache: bool = False,
            deepnorm: bool = False,
            pad_token_id=0,
            bos_token_id=1,
            eos_token_id=2,
            tie_word_embeddings=False,

            **kwargs,
    ):
        self.hidden_act = hidden_act
        self.hidden_size = hidden_size
        self.query_key_dim = query_key_dim
        self.initializer_range = initializer_range
        self.max_position_embeddings = max_position_embeddings
        self.num_attention_heads = num_attention_heads
        self.num_hidden_layers = num_hidden_layers
        self.rms_norm_eps = rms_norm_eps
        self.layernorm_eps = layernorm_eps
        self.retnorm = retnorm
        self.vocab_size = vocab_size
        self.v_factor = v_factor
        self.intermediate_k_select_scale = intermediate_k_select_scale
        self.intermediate_v_select_scale = intermediate_v_select_scale
        self.dense_block_layers = dense_block_layers
        self.dropout = dropout
        self.use_cache = use_cache
        self.deepnorm = deepnorm

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