configuration_ddllama.py

# coding=utf-8
# Copyright 2022 EleutherAI 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.

"""DDLLaMA model configuration"""

from transformers.configuration_utils import PretrainedConfig
from transformers.modeling_rope_utils import rope_config_validation

class DDLlamaConfig(PretrainedConfig):
    model_type = "ddllama"
    keys_to_ignore_at_inference = ["past_key_values"]
    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",
    }

    def __init__(
        self,
        vocab_size=128256,
        hidden_size=4096,
        intermediate_size=14336,
        num_hidden_layers=32,
        num_attention_heads=32,
        num_key_value_heads=8,
        router_layers=[16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31],
        router_reduction_factor=16,
        proj_reduction_factor=16,
        hidden_act="silu",
        max_position_embeddings=8192,
        initializer_range=0.02,
        rms_norm_eps=1e-5,
        use_cache=True,
        pad_token_id=None,
        bos_token_id=128000,
        eos_token_id=128009,
        pretraining_tp=1,
        tie_word_embeddings=False,
        rope_theta=500000.0,
        rope_scaling=None,
        attention_bias=False,
        attention_dropout=0.0,
        mlp_bias=False,
        head_dim=128,
        torch_dtype="bfloat16",
        **kwargs,
    ):
        self.vocab_size = vocab_size
        self.hidden_size = hidden_size
        self.intermediate_size = intermediate_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.router_layers = router_layers
        self.router_reduction_factor = router_reduction_factor
        self.proj_reduction_factor = proj_reduction_factor
        self.hidden_act = hidden_act
        self.max_position_embeddings = max_position_embeddings
        self.initializer_range = initializer_range
        self.rms_norm_eps = rms_norm_eps
        self.use_cache = use_cache
        self.pretraining_tp = pretraining_tp
        self.rope_theta = rope_theta
        self.rope_scaling = rope_scaling
        self.attention_bias = attention_bias
        self.attention_dropout = attention_dropout
        self.mlp_bias = mlp_bias
        self.head_dim = head_dim
        self.torch_dtype = torch_dtype
        
    
        # Validate the RoPE configuration
        rope_config_validation(self)

        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,
        )