configuration_pangu_moe.py

# coding=utf-8
# Copyright (c) Huawei Technologies Co., Ltd. 2025. All rights reserved.
# Copyright 2024 The Qwen team, Alibaba Group and 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.
""" PanguProMoE model configuration"""


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


logger = logging.get_logger(__name__)


class PanguProMoEConfig(PretrainedConfig):
   
    model_type = "PanguProMoE"
    _auto_class = "AutoConfig"

    def __init__(
        self,
        vocab_size=153376,
        hidden_size=4608,
        intermediate_size=10240,
        num_hidden_layers=50,
        num_attention_heads=64,
        num_key_value_heads=4,
        mlp_only_layers=[0,1,2,3],
        hidden_act="silu",
        max_position_embeddings=8192,
        initializer_range=0.02,
        rms_norm_eps=1e-5,
        use_cache=True,
        tie_word_embeddings=False,
        rope_theta=100000,
        moe_intermediate_size=1280,
        shared_expert_intermediate_size=2560,
        num_experts_per_tok=8,
        num_experts=80,
        norm_topk_prob=True,
        router_enable_expert_bias=True,
        output_router_logits=False,
        routed_scaling_factor=2.5,
        qk_nope_dim = 128,
        qk_rope_dim = 64,
        v_channels = 128,
        sandwich_norm=True,
        param_sink_number = 128,
        param_sink_with_value=True,
        **kwargs,
    ):
        self.vocab_size = vocab_size
        self.max_position_embeddings = max_position_embeddings
        self.hidden_size = hidden_size
        self.num_hidden_layers = num_hidden_layers
        self.num_attention_heads = num_attention_heads
        self.num_key_value_heads = num_key_value_heads
        self.hidden_act = hidden_act
        self.initializer_range = initializer_range
        self.rms_norm_eps = rms_norm_eps
        self.use_cache = use_cache
        self.rope_theta = rope_theta
        self.mlp_only_layers = mlp_only_layers
        self.intermediate_size = intermediate_size

        # MoE arguments
        self.moe_intermediate_size = moe_intermediate_size
        self.shared_expert_intermediate_size = shared_expert_intermediate_size
        self.num_experts_per_tok = num_experts_per_tok
        self.num_experts = num_experts
        self.norm_topk_prob = norm_topk_prob
        self.output_router_logits = output_router_logits
        self.router_enable_expert_bias = router_enable_expert_bias
        self.routed_scaling_factor = routed_scaling_factor
        self.qk_nope_dim = qk_nope_dim
        self.qk_rope_dim = qk_rope_dim
        self.v_channels = v_channels
        self.sandwich_norm = sandwich_norm
        self.param_sink_number = param_sink_number
        self.param_sink_with_value = param_sink_with_value

        super().__init__(
            tie_word_embeddings=tie_word_embeddings,
            **kwargs,
        )