mlp.py 9.89 KB
Newer Older
jixx's avatar
init  
jixx committed
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
import torch
import math
from torch import nn
from torch.nn import functional as F
from typing import Optional, Tuple
from text_generation_server.layers import TensorParallelEmbedding, FastLinear
from text_generation_server.layers.tensor_parallel import TensorParallelHead
from text_generation_server.utils.speculate import get_speculate


class MLPSpeculatorLayerNorm(nn.Module):
    """
    A L2 normalization implementation
    ...
    Args
    ----
    normalized_shape : int
        Dimensionality of input data (size of final tensor axis)
    elementwise_scale_weight : torch.Tensor
        learned scaling term after normalization?
    elementwise_shift_bias : torch.Tensor
        learned bias term after normalization?
    eps : float
        Safety term to prevent division by zero. Make sure the chosen value fits in the range of your encoding scheme (i.e. fp16 requires eps >= 6e-8).
    """

    def __init__(
        self,
        prefix,
        config,
        weights,
        eps=1e-06,
    ):
        super(MLPSpeculatorLayerNorm, self).__init__()
        self.weight = weights.get_tensor(f"{prefix}.weight")
        self.bias = weights.get_tensor(f"{prefix}.bias")
        self.eps = eps

    def forward(self, x):
        xf = x
        xf = xf * torch.rsqrt(xf.pow(2).mean(-1, keepdim=True) + self.eps)
        x = xf.type_as(x)
        x = self.weight * x
        x = x + self.bias
        return x


jixx's avatar
jixx committed
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
INV_SQRT2 = 2**-0.5


def simple_norm(x: torch.Tensor, eps=1e-06):
    xf = x
    xf = xf * torch.rsqrt(xf.pow(2).mean(-1, keepdim=True) + eps)
    x = xf.type_as(x)
    return x * INV_SQRT2


class MLPSpeculatorModelTied(torch.nn.Module):
    def __init__(self, config, prefix, weights):
        super().__init__()
        self.config = config
        self.n_predict = get_speculate()
        self.hidden_size = config.hidden_size

        self.emb = TensorParallelEmbedding(f"{prefix}.emb.0", weights)
        self.proj0 = FastLinear.load(
            config,
            prefix=f"{prefix}.proj.0",
            weights=weights,
            bias=False,
        )
        self.proj1 = FastLinear.load(
            config,
            prefix=f"{prefix}.proj.1",
            weights=weights,
            bias=False,
        )
        self.head = FastLinear.load(config, f"{prefix}.head.0", weights, bias=False)
        self.ln = MLPSpeculatorLayerNorm(
            prefix=f"{prefix}.ln.0",
            config=config,
            weights=weights,
        )

        # Weights ensure that state_0 accounts for 50% of state magnitude by final head in expectation
        self.state_weight = 0.5 ** (0.5 / self.n_predict) if self.n_predict > 0 else 1
        self.activation = nn.GELU()
        self.vsize = config.vocab_size
        self.inner_dim = config.speculator_config["inner_dim"]
        self.top_k_tokens_per_head = [1] * self.n_predict
        self.emb_weight = math.sqrt(1 - self.state_weight**2) * math.sqrt(
            self.inner_dim / 2
        )
        self.emb.weight *= self.emb_weight

    def forward(
        self,
        hidden_states: torch.Tensor,
        input_ids: torch.Tensor,
    ):
        top_k_tokens_per_head = self.top_k_tokens_per_head

        # k indicates # of candidates
        # h indicates # of generated tokens
        state = hidden_states
        b = state.size(0)
        ind = input_ids.unsqueeze(0)
        all_probs = torch.empty(
            b, self.n_predict, self.vsize, device=state.device
        )  # b k h v
        assert (
            len(top_k_tokens_per_head) == self.n_predict
        ), f"You must provide a topk number for each head ({self.n_predict} heads, {len(top_k_tokens_per_head)} provided)"
        for i in range(self.n_predict):
            # Project and predict
            z = self.emb(ind)
            # z = z.mul(self.emb_weight)  # b k d
            if i == 0:
                state = self.proj0(state) * self.state_weight + z
            else:
                state = self.proj1(state) * self.state_weight + z
            state = self.activation(self.ln(state))  # b k d
            probs = F.log_softmax(self.head(state), dim=-1)  # b k v
            _probs, preds = probs.topk(top_k_tokens_per_head[i], dim=-1)  # b k k'

            # Update candidate set with new predictions

            # Update distribution set with new logits
            all_probs[:, i] = probs.exp()

            # Update state, log_probs and ind for new predictions
            state = state.unsqueeze(2).expand(
                -1, -1, top_k_tokens_per_head[i], -1
            )  # b k k' d
            state = state.reshape(-1, b, state.size(3))  # b kk' d
            ind = preds.view(-1, b)  # b kk'

        speculative_logits = all_probs
        return speculative_logits


jixx's avatar
init  
jixx committed
142
143
144
145
146
147
class MLPSpeculatorModel(torch.nn.Module):
    def __init__(self, config, prefix, weights):
        super().__init__()
        self.config = config
        self.n_predict = get_speculate()
        self.hidden_size = config.hidden_size
jixx's avatar
jixx committed
148

jixx's avatar
init  
jixx committed
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
        self.emb = nn.ModuleList(
            [
                TensorParallelEmbedding(f"{prefix}.emb.{i}", weights)
                for i in range(self.n_predict)
            ]
        )
        self.proj = [
            FastLinear.load(
                config,
                prefix=f"{prefix}.proj.{i}",
                weights=weights,
                bias=False,
            )
            for i in range(self.n_predict)
        ]
        self.head = nn.ModuleList(
            [
                FastLinear.load(config, f"{prefix}.head.{i}", weights, bias=False)
                for i in range(self.n_predict)
            ]
        )
        self.ln = nn.ModuleList(
            [
                MLPSpeculatorLayerNorm(
                    prefix=f"{prefix}.ln.{i}",
                    config=config,
                    weights=weights,
                )
                for i in range(self.n_predict)
            ]
        )

        # Weights ensure that state_0 accounts for 50% of state magnitude by final head in expectation
jixx's avatar
jixx committed
182
        self.state_weight = 0.5 ** (0.5 / self.n_predict) if self.n_predict > 0 else 1
jixx's avatar
init  
jixx committed
183
184
185
186
        self.activation = nn.GELU()
        self.vsize = config.vocab_size
        self.inner_dim = config.speculator_config["inner_dim"]
        self.top_k_tokens_per_head = [1] * self.n_predict
jixx's avatar
jixx committed
187
188
189
190
        self.emb_weight = math.sqrt(1 - self.state_weight**2) * math.sqrt(
            self.inner_dim / 2
        )
        self.emb.weight *= self.emb_weight
jixx's avatar
init  
jixx committed
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212

    def forward(
        self,
        hidden_states: torch.Tensor,
        input_ids: torch.Tensor,
    ):
        top_k_tokens_per_head = self.top_k_tokens_per_head

        # k indicates # of candidates
        # h indicates # of generated tokens
        state = hidden_states
        b = state.size(0)
        ind = input_ids.unsqueeze(0)
        all_probs = torch.empty(
            b, self.n_predict, self.vsize, device=state.device
        )  # b k h v
        assert (
            len(top_k_tokens_per_head) == self.n_predict
        ), f"You must provide a topk number for each head ({self.n_predict} heads, {len(top_k_tokens_per_head)} provided)"
        for i in range(self.n_predict):
            # Project and predict
            z = self.emb[i](ind)
jixx's avatar
jixx committed
213
            # z = z.mul(self.emb_weight)  # b k d
jixx's avatar
init  
jixx committed
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
            state = self.proj[i](state) * self.state_weight + z
            state = self.activation(self.ln[i](state))  # b k d
            probs = F.log_softmax(self.head[i](state), dim=-1)  # b k v
            _probs, preds = probs.topk(top_k_tokens_per_head[i], dim=-1)  # b k k'

            # Update candidate set with new predictions

            # Update distribution set with new logits
            all_probs[:, i] = probs.exp()

            # Update state, log_probs and ind for new predictions
            state = state.unsqueeze(2).expand(
                -1, -1, top_k_tokens_per_head[i], -1
            )  # b k k' d
            state = state.reshape(-1, b, state.size(3))  # b kk' d
            ind = preds.view(-1, b)  # b kk'

        speculative_logits = all_probs
        return speculative_logits


class MLPSpeculatorHead(nn.Module):
jixx's avatar
jixx committed
236
    def __init__(self, lm_head, mlp_speculator, scale_input: bool):
jixx's avatar
init  
jixx committed
237
238
239
        super().__init__()
        self.lm_head = lm_head
        self.mlp_speculator = mlp_speculator
jixx's avatar
jixx committed
240
        self.scale_input = scale_input
jixx's avatar
init  
jixx committed
241
242
243
244
245
246
247
248
249
250

    def forward(
        self, input: torch.Tensor
    ) -> Tuple[torch.Tensor, Optional[torch.Tensor]]:
        logits = self.lm_head(input)
        # If we have too many tokens, we skip speculative logits
        if input.shape[0] > 128:
            return logits, None

        input_ids = logits.argmax(dim=-1)
jixx's avatar
jixx committed
251
252
        if self.scale_input:
            input = simple_norm(input)
jixx's avatar
init  
jixx committed
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
        speculative_logits = self.mlp_speculator(input, input_ids)
        return logits, speculative_logits

    @staticmethod
    def load(config, prefix: str, weights):
        from pathlib import Path
        from safetensors import safe_open

        speculator_path = config.speculator["path"]

        for fname in config.speculator["model_paths"]:
            filename = str(Path(speculator_path) / fname)
            routing = weights.routing
            with safe_open(filename, framework="pytorch") as f:
                for k in f.keys():
                    if k in routing and routing[k] != filename:
                        raise RuntimeError(
                            f"Key {k} was found in multiple files: {filename} and {routing[k]}"
                        )
                    routing[k] = filename

jixx's avatar
jixx committed
274
275
276
277
278
279
280
        tie_weights = config.speculator_config.get("tie_weights", False)
        if tie_weights:
            mlp_speculator = MLPSpeculatorModelTied(config, "speculator", weights)
        else:
            mlp_speculator = MLPSpeculatorModel(config, "speculator", weights)
        # This is used in https://huggingface.co/ibm-fms/llama3-70b-accelerator
        scale_input = config.speculator_config.get("scale_input", False)
jixx's avatar
init  
jixx committed
281
        lm_head = TensorParallelHead.load(config, prefix, weights)
jixx's avatar
jixx committed
282
        return MLPSpeculatorHead(lm_head, mlp_speculator, scale_input)