configuration_rwkv07f.py

# coding=utf-8
# 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.
"""RWKV07DQwen3 model configuration"""

#Never gonna give you up

from transformers.configuration_utils import PretrainedConfig, layer_type_validation
from transformers.modeling_rope_utils import rope_config_validation
from transformers.utils import logging


from transformers.configuration_utils import PretrainedConfig, layer_type_validation
#from transformers.modeling_rope_utils import RopeParameters
from typing import Optional, TypedDict
#from transformers.modeling_rope_utils import RopeParameters
class RopeParameters(TypedDict):
    """
    Args:
        rope_theta (`float`):
            The base period of the RoPE embeddings.
        rope_type (`str`, *optional*, defaults to "default"):
            The sub-variant of RoPE to use. Can be one of ['default', 'linear', 'dynamic', 'yarn', 'longrope',
            'llama3'], with 'default' being the original RoPE implementation.
        factor (`float`, *optional*):
            Used with all rope types except 'default'. The scaling factor to apply to the RoPE embeddings. In
            most scaling types, a `factor` of x will enable the model to handle sequences of length x *
            original maximum pre-trained length.
        original_max_position_embeddings (`int`, *optional*):
            Used with 'dynamic', 'longrope' and 'llama3'. The original max position embeddings used during
            pretraining.
        attention_factor (`float`, *optional*):
            Used with 'yarn' and 'longrope'. The scaling factor to be applied on the attention
            computation. If unspecified, it defaults to value recommended by the implementation, using the
            `factor` field to infer the suggested value.
        beta_fast (`float`, *optional*):
            Only used with 'yarn'. Parameter to set the boundary for extrapolation (only) in the linear
            ramp function. If unspecified, it defaults to 32.
        beta_slow (`float`, *optional*):
            Only used with 'yarn'. Parameter to set the boundary for interpolation (only) in the linear
            ramp function. If unspecified, it defaults to 1.
        short_factor (`list[float]`, *optional*):
            Only used with 'longrope'. The scaling factor to be applied to short contexts (<
            `original_max_position_embeddings`). Must be a list of numbers with the same length as the hidden
            size divided by the number of attention heads divided by 2
        long_factor (`list[float]`, *optional*):
            Only used with 'longrope'. The scaling factor to be applied to long contexts (<
            `original_max_position_embeddings`). Must be a list of numbers with the same length as the hidden
            size divided by the number of attention heads divided by 2
        low_freq_factor (`float`, *optional*):
            Only used with 'llama3'. Scaling factor applied to low frequency components of the RoPE
        high_freq_factor (`float`, *optional*):
            Only used with 'llama3'. Scaling factor applied to high frequency components of the RoPE
    """

    rope_theta: float
    rope_type: Optional[str]
    factor: Optional[float]
    original_max_position_embeddings: Optional[int]
    attention_factor: Optional[float]
    beta_fast: Optional[float]
    beta_slow: Optional[float]
    short_factor: Optional[list[float]]
    long_factor: Optional[list[float]]
    low_freq_factor: Optional[float]
    high_freq_factor: Optional[float]
logger = logging.get_logger(__name__)


class RWKV07FConfig(PretrainedConfig):
    r"""
    This is the configuration class to store the configuration of a [`RWKV07BModel`]. It is used to instantiate a
    RWKV079Qwen3 model according to the specified arguments, defining the model architecture. Instantiating a configuration
    with the defaults will yield a similar configuration to that of
    Qwen3-7B-beta [Qwen/Qwen3-7B-beta](https://huggingface.co/Qwen/Qwen3-7B-beta).

    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 151936):
            Vocabulary size of the RWKV079Qwen3 model. Defines the number of different tokens that can be represented by the
            `inputs_ids` passed when calling [`RWKV07BModel`]
        hidden_size (`int`, *optional*, defaults to 4096):
            Dimension of the hidden representations.
        intermediate_size (`int`, *optional*, defaults to 22016):
            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*, defaults to 32):
            This is the number of key_value heads that should be used to implement Grouped Query Attention. If
            `num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if
            `num_key_value_heads=1` the model will use Multi Query Attention (MQA) otherwise GQA is used. When
            converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed
            by meanpooling all the original heads within that group. For more details checkout [this
            paper](https://arxiv.org/pdf/2305.13245.pdf). If it is not specified, will default to `32`.
        lora_rank_decay (`int`, *optional*):
            The rank of the lora used to generate decay.
        lora_rank_iclr (`int`, *optional*):
            The rank of the lora used to generate the in-context learning rate.
        lora_rank_value_residual_mix (`int`, *optional*):
            The rank of the lora used to generate the value residual mix amount.
        lora_rank_value_gate (`int`, *optional*):
            The rank of the lora used to generate the gate.
        hidden_act (`str` or `function`, *optional*, defaults to `"silu"`):
            The non-linear activation function (function or string) in the decoder.
        max_position_embeddings (`int`, *optional*, defaults to 32768):
            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 `True`):
            Whether or not the model should return the last key/values attentions (not used by all models). Only
            relevant if `config.is_decoder=True`.
        tie_word_embeddings (`bool`, *optional*, defaults to `False`):
            Whether the model's input and output word embeddings should be tied.
        rope_theta (`float`, *optional*, defaults to 10000.0):
            The base period of the RoPE embeddings.
        rope_scaling (`Dict`, *optional*):
            Dictionary containing the scaling configuration for the RoPE embeddings. NOTE: if you apply new rope type
            and you expect the model to work on longer `max_position_embeddings`, we recommend you to update this value
            accordingly.
            Expected contents:
                `rope_type` (`str`):
                    The sub-variant of RoPE to use. Can be one of ['default', 'linear', 'dynamic', 'yarn', 'longrope',
                    'llama3'], with 'default' being the original RoPE implementation.
                `factor` (`float`, *optional*):
                    Used with all rope types except 'default'. The scaling factor to apply to the RoPE embeddings. In
                    most scaling types, a `factor` of x will enable the model to handle sequences of length x *
                    original maximum pre-trained length.
                `original_max_position_embeddings` (`int`, *optional*):
                    Used with 'dynamic', 'longrope' and 'llama3'. The original max position embeddings used during
                    pretraining.
                `attention_factor` (`float`, *optional*):
                    Used with 'yarn' and 'longrope'. The scaling factor to be applied on the attention
                    computation. If unspecified, it defaults to value recommended by the implementation, using the
                    `factor` field to infer the suggested value.
                `beta_fast` (`float`, *optional*):
                    Only used with 'yarn'. Parameter to set the boundary for extrapolation (only) in the linear
                    ramp function. If unspecified, it defaults to 32.
                `beta_slow` (`float`, *optional*):
                    Only used with 'yarn'. Parameter to set the boundary for interpolation (only) in the linear
                    ramp function. If unspecified, it defaults to 1.
                `short_factor` (`List[float]`, *optional*):
                    Only used with 'longrope'. The scaling factor to be applied to short contexts (<
                    `original_max_position_embeddings`). Must be a list of numbers with the same length as the hidden
                    size divided by the number of attention heads divided by 2
                `long_factor` (`List[float]`, *optional*):
                    Only used with 'longrope'. The scaling factor to be applied to long contexts (<
                    `original_max_position_embeddings`). Must be a list of numbers with the same length as the hidden
                    size divided by the number of attention heads divided by 2
                `low_freq_factor` (`float`, *optional*):
                    Only used with 'llama3'. Scaling factor applied to low frequency components of the RoPE
                `high_freq_factor` (`float`, *optional*):
                    Only used with 'llama3'. Scaling factor applied to high frequency components of the RoPE
        use_sliding_window (`bool`, *optional*, defaults to `False`):
            Whether to use sliding window attention.
        sliding_window (`int`, *optional*, defaults to 4096):
            Sliding window attention (SWA) window size. If not specified, will default to `4096`.
        max_window_layers (`int`, *optional*, defaults to 28):
            The number of layers that use SWA (Sliding Window Attention). The bottom layers use SWA while the top use full attention.
        attention_dropout (`float`, *optional*, defaults to 0.0):
            The dropout ratio for the attention probabilities.

    ```python
    >>> from transformers import RWKV07BModel, RWKV079Qwen3Config

    >>> # Initializing a RWKV079Qwen3 style configuration
    >>> configuration = RWKV079Qwen3Config()

    >>> # Initializing a model from the RWKV079Qwen3-7B style configuration
    >>> model = RWKV07BModel(configuration)

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

    model_type = "rwkv07f_moe"

    keys_to_ignore_at_inference = ["past_key_values"]
    base_model_tp_plan = {
        "layers.*.self_attn.o_proj": "rowwise",
        "layers.*.mlp.experts.gate_up_proj": "local_rowwise",
        "layers.*.mlp.experts.down_proj": "local_rowwise",
        "layers.*.mlp.experts": "gather",
        "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"]),
    }
    attribute_map = {
        "num_local_experts": "n_routed_experts",
    }

    def __init__(
        self,
        lora_rank_tokenshift=None,
        lora_rank_decay=None,
        lora_rank_iclr=None,
        lora_rank_value_residual_mix=None,
        lora_rank_value_key_mix=None,
        lora_rank_gate=None,

        vocab_size: int | None = 154880,
        hidden_size: int | None = 2048,
        intermediate_size: int | None = 10240,
        moe_intermediate_size: int | None = 1536,
        num_hidden_layers: int | None = 47,
        num_attention_heads: int | None = 20,
        num_key_value_heads: int | None = 20,
        n_shared_experts: int | None = 1,
        n_routed_experts: int | None = 64,
        routed_scaling_factor: float | None = 1.8,
        kv_lora_rank: int | None = 512,
        q_lora_rank: int | None = 768,
        qk_rope_head_dim: int | None = 64,
        v_head_dim: int | None = 256,
        qk_nope_head_dim: int | None = 192,
        n_group: int | None = 1,
        topk_group: int | None = 1,
        num_experts_per_tok: int | None = 4,
        norm_topk_prob: bool | None = True,
        hidden_act: str | None = "silu",
        max_position_embeddings: int | None = 202752,
        initializer_range: float | None = 0.02,
        rms_norm_eps: int | None = 1e-5,
        use_cache: bool | None = True,
        pad_token_id: int | None = None,
        bos_token_id: int | None = 0,
        eos_token_id: int | None = 1,
        pretraining_tp: int | None = 1,
        tie_word_embeddings: bool | None = False,
        rope_parameters: RopeParameters | dict[str, RopeParameters] | None = None,
        rope_interleave: bool | None = True,
        mlp_layer_types=None,
        attention_bias: bool | None = False,
        attention_dropout: float | None = 0.0,
        
        **kwargs,
    ):


        self.num_key_value_heads = num_key_value_heads
        self.lora_rank_tokenshift = lora_rank_tokenshift
        self.lora_rank_decay = lora_rank_decay
        self.lora_rank_iclr = lora_rank_iclr
        self.lora_rank_value_residual_mix = lora_rank_value_residual_mix
        self.lora_rank_gate = lora_rank_gate

        self.vocab_size = vocab_size
        self.max_position_embeddings = max_position_embeddings
        self.hidden_size = hidden_size
        self.intermediate_size = intermediate_size
        self.num_hidden_layers = num_hidden_layers

        # Default to MoE from the second layer and on
        self.mlp_layer_types = mlp_layer_types
        if self.mlp_layer_types is None:
            self.mlp_layer_types = ["dense"] + ["sparse"] * (self.num_hidden_layers - 1)
        layer_type_validation(self.mlp_layer_types, self.num_hidden_layers, attention=False)

        self.layer_types = None
        self.sliding_window = None
        if self.layer_types is None:
            self.layer_types = [
                "sliding_attention"
                if self.sliding_window is not None and i >= self.max_window_layers
                else "full_attention"
                for i in range(self.num_hidden_layers)
            ]

        self.moe_intermediate_size = moe_intermediate_size
        self.num_attention_heads = num_attention_heads
        self.n_shared_experts = n_shared_experts
        self.n_routed_experts = n_routed_experts
        self.routed_scaling_factor = routed_scaling_factor
        self.kv_lora_rank = kv_lora_rank
        self.q_lora_rank = q_lora_rank
        self.qk_rope_head_dim = qk_rope_head_dim
        self.v_head_dim = v_head_dim
        self.qk_nope_head_dim = qk_nope_head_dim
        self.qk_head_dim = qk_nope_head_dim + qk_rope_head_dim
        self.head_dim = qk_rope_head_dim
        self.n_group = n_group
        self.topk_group = topk_group
        self.num_experts_per_tok = num_experts_per_tok
        self.norm_topk_prob = norm_topk_prob
        self.rope_interleave = rope_interleave
        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.pretraining_tp = pretraining_tp
        self.use_cache = use_cache
        self.attention_bias = attention_bias
        self.attention_dropout = attention_dropout
        self.rope_parameters = rope_parameters
        self.pad_token_id = pad_token_id
        self.bos_token_id = bos_token_id
        self.eos_token_id = eos_token_id
        self.tie_word_embeddings = tie_word_embeddings

        super().__init__(**kwargs)

__all__ = ["RWKV07FConfig"]