vllm.model_executor.models.mlp_speculator
MLPSpeculator
¶
Bases: Module
An implementation of the speculative models introduced in "Accelerating Production LLMs with Combined Token/Embedding Speculators" https://arxiv.org/pdf/2404.19124
Trained speculators of this type are available on HF hub at: https://huggingface.co/ibm-ai-platform and https://huggingface.co/ibm-granite
Source code in vllm/model_executor/models/mlp_speculator.py
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 |
|
logits_processor
instance-attribute
¶
logits_processor = LogitsProcessor(
vocab_size, vocab_size, 1.0
)
proj
instance-attribute
¶
proj = ModuleList(
[proj_first] + [proj_tied] * max_speculative_tokens - 1
)
__init__
¶
__init__(
*, vllm_config: VllmConfig, prefix: str = ""
) -> None
Source code in vllm/model_executor/models/mlp_speculator.py
generate_proposals
¶
generate_proposals(
input_ids: Tensor,
previous_hidden_states: Tensor,
num_predict_tokens: int,
sampling_metadata: SamplingMetadata,
) -> list[SamplerOutput]
Source code in vllm/model_executor/models/mlp_speculator.py
load_weights
¶
Source code in vllm/model_executor/models/mlp_speculator.py
MLPSpeculatorLayerNorm
¶
Bases: Module
A L2 normalization implementation ... Args
normalized_shape : int Dimensionality of input data (size of final tensor axis) 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). elementwise_scale_and_shift : bool Include a learned scaling and shift term after normalization.