Get the number of parameters in the embedding layer. Args: shared_embedding (bool, optional): whether the output embedding \ shares weights with the input embedding. Defaults to True. Returns: int: the number of parameters in the embedding l
(self, shared_embedding: bool = True)
| 242 | return pivot |
| 243 | |
| 244 | def get_num_params_embedding(self, shared_embedding: bool = True) -> int: |
| 245 | """Get the number of parameters in the embedding layer. |
| 246 | |
| 247 | Args: |
| 248 | shared_embedding (bool, optional): whether the output embedding \ |
| 249 | shares weights with the input embedding. Defaults to True. |
| 250 | |
| 251 | Returns: |
| 252 | int: the number of parameters in the embedding layer |
| 253 | """ |
| 254 | num_params_input_embedding = (self.model_config.hidden_dim * |
| 255 | self.model_config.vocab_size) |
| 256 | num_params_output_embedding = (self.model_config.hidden_dim * |
| 257 | self.model_config.vocab_size |
| 258 | if not shared_embedding else 0) |
| 259 | return num_params_input_embedding + num_params_output_embedding |
| 260 | |
| 261 | def get_num_params_per_layer_attn(self) -> int: |
| 262 | """Get the number of parameters in the attention linear layers, including the |
no outgoing calls
no test coverage detected