(self, config: Gemma2Config)
| 653 | """ |
| 654 | |
| 655 | def __init__(self, config: Gemma2Config): |
| 656 | super().__init__(config) |
| 657 | self.padding_idx = config.pad_token_id |
| 658 | self.vocab_size = config.vocab_size |
| 659 | |
| 660 | self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx) |
| 661 | self.layers = nn.ModuleList( |
| 662 | [Gemma2DecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)] |
| 663 | ) |
| 664 | self.norm = Gemma2RMSNorm(config.hidden_size, eps=config.rms_norm_eps) |
| 665 | |
| 666 | self.gradient_checkpointing = False |
| 667 | if getattr(config, "pretraining_tp", 1) != 1: |
| 668 | logger.warn("`pretraining_tp` is deprecated, please use `model.tensor_parallel` instead.") |
| 669 | |
| 670 | # Initialize weights and apply final processing |
| 671 | self.post_init() |
| 672 | |
| 673 | def get_input_embeddings(self): |
| 674 | return self.embed_tokens |
nothing calls this directly
no test coverage detected