File size: 8,205 Bytes
a2e5c4f
866d907
a2e5c4f
 
 
 
 
 
 
 
 
 
 
 
6caf34c
a2e5c4f
6caf34c
 
a2e5c4f
 
 
 
6caf34c
866d907
d4ebf23
 
 
6caf34c
 
 
 
 
 
 
 
 
d4ebf23
6caf34c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d4ebf23
6caf34c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
687dc4c
6caf34c
 
 
1697717
687dc4c
6caf34c
 
 
 
 
 
 
 
 
 
 
d4ebf23
6caf34c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1697717
a2c4034
a6efa5f
6caf34c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d4ebf23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
# coding=utf-8
# Copyright 2023 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.
""" Quiet model configuration"""

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


logger = logging.get_logger(__name__)

QUIET_PRETRAINED_CONFIG_ARCHIVE_MAP = {
    "cognitivecomputations/Quiet-STaR-Base": "https://huggingface.co/quietai/Quiet-7B-v0.1/resolve/main/config.json",
}


class QuietConfig(PretrainedConfig):
    r"""
    This is the configuration class to store the configuration of a [`QuietModel`]. It is used to instantiate an
    Quiet 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 Quiet-7B-v0.1 or Quiet-7B-Instruct-v0.1.
    [quietai/Quiet-7B-v0.1](https://huggingface.co/quietai/Quiet-7B-v0.1)
    [quietai/Quiet-7B-Instruct-v0.1](https://huggingface.co/quietai/Quiet-7B-Instruct-v0.1)
    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 32000):
            Vocabulary size of the Quiet model. Defines the number of different tokens that can be represented by the
            `inputs_ids` passed when calling [`QuietModel`]
        hidden_size (`int`, *optional*, defaults to 4096):
            Dimension of the hidden representations.
        intermediate_size (`int`, *optional*, defaults to 14336):
            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 8):
            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 `8`.
        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 `4096*32`):
            The maximum sequence length that this model might ever be used with. Quiet's sliding window attention
            allows sequence of up to 4096*32 tokens.
        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`.
        pad_token_id (`int`, *optional*):
            The id of the padding token.
        bos_token_id (`int`, *optional*, defaults to 1):
            The id of the "beginning-of-sequence" token.
        eos_token_id (`int`, *optional*, defaults to 2):
            The id of the "end-of-sequence" token.
        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.
        sliding_window (`int`, *optional*, defaults to 4096):
            Sliding window attention window size. If not specified, will default to `4096`.
        attention_dropout (`float`, *optional*, defaults to 0.0):
            The dropout ratio for the attention probabilities.
    ```python
    >>> from transformers import QuietModel, QuietConfig
    >>> # Initializing a Quiet 7B style configuration
    >>> configuration = QuietConfig()
    >>> # Initializing a model from the Quiet 7B style configuration
    >>> model = QuietModel(configuration)
    >>> # Accessing the model configuration
    >>> configuration = model.config
    ```"""

    model_type = "quiet"
    keys_to_ignore_at_inference = ["past_key_values"]

    def __init__(
        self,
        vocab_size=32000,
        hidden_size=4096,
        intermediate_size=14336,
        num_hidden_layers=32,
        num_attention_heads=32,
        num_key_value_heads=8,
        hidden_act="silu",
        max_position_embeddings=4096 * 32,
        initializer_range=0.02,
        rms_norm_eps=1e-6,
        use_cache=True,
        pad_token_id=None,
        bos_token_id=1,
        eos_token_id=2,
        tie_word_embeddings=False,
        rope_theta=10000.0,
        complexity_factor = 0.5,
        sliding_window=4096,
        attention_dropout=0.0,
        max_thoughts=16,
        max_time=None,
        max_temperature=10,
        merged_talk_heads=True,
        merged_lm_and_talk_heads=False,
        merged_lm_and_think_heads=True,
        use_concat_talk_head=True,
        use_shallow_think=True,
        use_shallow_talk=False,
        use_complex_think_head=False,
        use_complex_talk_head=True,
        use_weighted_talk_head=True,
        hidden_dropout_prob=0.0, 
        **kwargs,
    ):
        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
        self.num_attention_heads = num_attention_heads
        self.sliding_window = sliding_window

        # for backward compatibility
        if num_key_value_heads is None:
            num_key_value_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.attention_dropout = attention_dropout
        self.max_thoughts = max_thoughts
        self.max_time = max_time
        self.complexity_factor = complexity_factor
        self.max_temperature = max_temperature
        self.merged_talk_heads = merged_talk_heads
        self.merged_lm_and_talk_heads = merged_lm_and_talk_heads
        self.merged_lm_and_think_heads = merged_lm_and_think_heads
        self.use_concat_talk_head = use_concat_talk_head
        self.use_shallow_think = use_shallow_think
        self.use_shallow_talk = use_shallow_talk
        self.use_complex_think_head = use_complex_think_head
        self.use_complex_talk_head = use_complex_talk_head
        self.use_weighted_talk_head = use_weighted_talk_head
        self.hidden_dropout_prob = hidden_dropout_prob

        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,
        )