Deci
/

File size: 12,339 Bytes
7a2815a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
# coding=utf-8
# Copyright and license here
""" PyTorch DeciCoder model."""
import math
from typing import Optional, Tuple

import torch
import torch.nn.functional as F
import torch.utils.checkpoint
from torch import nn
from transformers.models.llama.modeling_llama import LlamaMLP, LlamaRMSNorm, LlamaAttention, apply_rotary_pos_emb, \
    repeat_kv, LlamaPreTrainedModel, LLAMA_START_DOCSTRING, LlamaDecoderLayer, LlamaForCausalLM, LlamaModel
from transformers.utils import add_start_docstrings

from .configuration_decicoder import DeciCoderConfig

_CONFIG_FOR_DOC = "DeciCoderConfig"


class DeciCoderAttention(LlamaAttention):
    """Multi-headed attention from 'Attention Is All You Need' paper"""

    def __init__(self, config: DeciCoderConfig):
        nn.Module.__init__(self)
        self.config = config
        self.hidden_size = config.hidden_size
        self.num_heads = config.num_attention_heads
        self.head_dim = self.hidden_size // self.num_heads
        self.num_key_value_heads = config.num_key_value_heads
        self.num_key_value_groups = self.num_heads // self.num_key_value_heads
        self.pretraining_tp = config.pretraining_tp
        self.max_position_embeddings = config.max_position_embeddings

        if (self.head_dim * self.num_heads) != self.hidden_size:
            raise ValueError(
                f"hidden_size must be divisible by num_heads (got `hidden_size`: {self.hidden_size}"
                f" and `num_heads`: {self.num_heads})."
            )
        self.q_proj = nn.Linear(self.hidden_size, self.num_heads * self.head_dim, bias=False)
        self.k_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
        self.v_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
        self.o_proj = nn.Linear(self.num_heads * self.head_dim, self.hidden_size, bias=False)

        self.naive_attention_prefill = config.naive_attention_prefill
        self.naive_attention_decode_batched = config.naive_attention_decode_batched
        self.naive_attention_decode_single = config.naive_attention_decode_single
        self._init_rope()

    def forward(
            self,
            hidden_states: torch.Tensor,
            attention_mask: Optional[torch.Tensor] = None,
            position_ids: Optional[torch.LongTensor] = None,
            past_key_value: Optional[Tuple[torch.Tensor]] = None,
            output_attentions: bool = False,
            use_cache: bool = False,
    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
        bsz, q_len, _ = hidden_states.size()
        if past_key_value is None:
            is_decode = False
        else:
            is_decode = True
        if self.pretraining_tp > 1:
            key_value_slicing = (self.num_key_value_heads * self.head_dim) // self.pretraining_tp
            query_slices = self.q_proj.weight.split((self.num_heads * self.head_dim) // self.pretraining_tp, dim=0)
            key_slices = self.k_proj.weight.split(key_value_slicing, dim=0)
            value_slices = self.v_proj.weight.split(key_value_slicing, dim=0)

            query_states = [F.linear(hidden_states, query_slices[i]) for i in range(self.pretraining_tp)]
            query_states = torch.cat(query_states, dim=-1)

            key_states = [F.linear(hidden_states, key_slices[i]) for i in range(self.pretraining_tp)]
            key_states = torch.cat(key_states, dim=-1)

            value_states = [F.linear(hidden_states, value_slices[i]) for i in range(self.pretraining_tp)]
            value_states = torch.cat(value_states, dim=-1)

        else:
            query_states = self.q_proj(hidden_states)
            key_states = self.k_proj(hidden_states)
            value_states = self.v_proj(hidden_states)

        query_states = query_states.view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
        key_states = key_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
        value_states = value_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)

        kv_seq_len = key_states.shape[-2]
        if past_key_value is not None:
            kv_seq_len += past_key_value[0].shape[-2]
        cos, sin = self.rotary_emb(value_states, seq_len=kv_seq_len)

        query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, position_ids)

        if past_key_value is not None:
            # reuse k, v, self_attention
            key_states = torch.cat([past_key_value[0], key_states], dim=2)
            value_states = torch.cat([past_key_value[1], value_states], dim=2)

        past_key_value = (key_states, value_states) if use_cache else None

        # repeat k/v heads if n_kv_heads < n_heads
        if is_decode:
            query_states = query_states.view(bsz, self.num_key_value_heads, self.num_key_value_groups, self.head_dim)
            if self.naive_attention_decode_batched and bsz > 1 or self.naive_attention_decode_single and bsz == 1:
                attn_weights = (query_states @ key_states.transpose(-2, -1)) / math.sqrt(key_states.size(-1))
                attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query_states.dtype)
                if attention_mask is not None:
                    if attention_mask.size() != (bsz, 1, q_len, kv_seq_len):
                        raise ValueError(
                            f"Attention mask should be of size {(bsz, 1, q_len, kv_seq_len)}, but is {attention_mask.size()}"
                        )
                    attn_weights = attn_weights + attention_mask

                attn_output = torch.matmul(attn_weights, value_states)
            else:
                attn_output = F.scaled_dot_product_attention(query_states, key_states, value_states, is_causal=False,
                                                             dropout_p=0.0)
            attn_output = attn_output.contiguous().view(bsz, q_len, self.hidden_size)

        else:
            key_states = repeat_kv(key_states, self.num_key_value_groups)
            value_states = repeat_kv(value_states, self.num_key_value_groups)

            if not self.naive_attention_prefill:
                attn_output = F.scaled_dot_product_attention(query_states, key_states, value_states, is_causal=True,
                                                             dropout_p=0.0)
            else:
                attn_weights = torch.matmul(query_states, key_states.transpose(2, 3)) / math.sqrt(self.head_dim)
                # attn_weights = (query_states @ key_states.transpose(-2, -1)) / math.sqrt(key_states.size(-1))
                if attn_weights.size() != (bsz, self.num_heads, q_len, kv_seq_len):
                    raise ValueError(
                        f"Attention weights should be of size {(bsz, self.num_heads, q_len, kv_seq_len)}, but is"
                        f" {attn_weights.size()}"
                    )

                if attention_mask is not None:
                    if attention_mask.size() != (bsz, 1, q_len, kv_seq_len):
                        raise ValueError(
                            f"Attention mask should be of size {(bsz, 1, q_len, kv_seq_len)}, but is {attention_mask.size()}"
                        )
                    attn_weights = attn_weights + attention_mask

                # upcast attention to fp32
                attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query_states.dtype)
                attn_output = torch.matmul(attn_weights, value_states)

            if attn_output.size() != (bsz, self.num_heads, q_len, self.head_dim):
                raise ValueError(
                    f"`attn_output` should be of size {(bsz, self.num_heads, q_len, self.head_dim)}, but is"
                    f" {attn_output.size()}"
                )

            attn_output = attn_output.transpose(1, 2).contiguous().view(bsz, q_len, self.hidden_size)
            # attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)

        if self.pretraining_tp > 1:
            attn_output = attn_output.split(self.hidden_size // self.pretraining_tp, dim=2)
            o_proj_slices = self.o_proj.weight.split(self.hidden_size // self.pretraining_tp, dim=1)
            attn_output = sum([F.linear(attn_output[i], o_proj_slices[i]) for i in range(self.pretraining_tp)])
        else:
            attn_output = self.o_proj(attn_output)

        if not output_attentions:
            attn_weights = None

        return attn_output, attn_weights, past_key_value


class DeciCoderDecoderLayer(LlamaDecoderLayer):
    def __init__(self, config: DeciCoderConfig):
        nn.Module.__init__(self)
        self.hidden_size = config.hidden_size
        self.self_attn = DeciCoderAttention(config=config)
        self.mlp = LlamaMLP(config)
        self.input_layernorm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.post_attention_layernorm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)


@add_start_docstrings(
    "The bare DeciCoder Model outputting raw hidden-states without any specific head on top.",
    LLAMA_START_DOCSTRING,
)
class DeciCoderPreTrainedModel(LlamaPreTrainedModel):
    config_class = DeciCoderConfig
    _no_split_modules = ["DeciCoderDecoderLayer"]
    _keys_to_ignore_on_load_missing = ["self_attn.rotary_emb.inv_freq"]


@add_start_docstrings(
    "The bare DeciCoder Model outputting raw hidden-states without any specific head on top.",
    LLAMA_START_DOCSTRING,
)
class DeciCoderModel(LlamaModel, DeciCoderPreTrainedModel):
    """
    Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`DeciCoderDecoderLayer`]

    Args:
        config: DeciCoderConfig
    """

    def __init__(self, config: DeciCoderConfig):
        DeciCoderPreTrainedModel.__init__(self, config)
        self.padding_idx = config.pad_token_id
        self.vocab_size = config.vocab_size

        self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
        self.layers = nn.ModuleList([DeciCoderDecoderLayer(config) for _ in range(config.num_hidden_layers)])
        self.norm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)

        self.gradient_checkpointing = False
        # Initialize weights and apply final processing
        self.post_init()

    def _prepare_decoder_attention_mask(self, attention_mask, input_shape, inputs_embeds, past_key_values_length):
        self._validate_config_supports_attention_mask(attention_mask, input_shape, past_key_values_length)
        return LlamaModel._prepare_decoder_attention_mask(
            self, attention_mask, input_shape, inputs_embeds, past_key_values_length)

    def _validate_config_supports_attention_mask(self, attention_mask, input_shape, past_key_values_length):
        is_decode = past_key_values_length > 0
        if not torch.all(torch.eq(attention_mask, 1)).item():
            if is_decode:
                if input_shape[0] == 1 and not self.config.naive_attention_decode_single:
                    raise ValueError(
                        "For support of custom attention masks please set naive_attention_decode_single to True in the "
                        "config")
                elif input_shape[0] > 1 and not self.config.naive_attention_decode_batched:
                    raise ValueError(
                        "For support of custom attention masks please set naive_attention_decode_batched to True in the"
                        "config")
            else:
                if not self.config.naive_attention_prefill:
                    raise ValueError("For support of custom attention masks please set naive_attention_prefill to "
                                     "True in the config")


class DeciCoderForCausalLM(LlamaForCausalLM, DeciCoderPreTrainedModel):
    def __init__(self, config):
        DeciCoderPreTrainedModel.__init__(self, config)
        self.model = DeciCoderModel(config)
        self.pretraining_tp = config.pretraining_tp
        self.vocab_size = config.vocab_size
        self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)

        # Initialize weights and apply final processing
        self.post_init()