| from transformers import PretrainedConfig | |
| class ShivikM3Config(PretrainedConfig): | |
| model_type = "shivik-m3" | |
| def __init__( | |
| self, | |
| vocab_size=200018, | |
| hidden_size=2048, | |
| num_hidden_layers=28, | |
| num_attention_heads=32, | |
| intermediate_size=7168, | |
| kv_head_split_layer=14, | |
| num_kv_heads=8, | |
| num_kv_heads_high=32, | |
| tie_word_embeddings=True, | |
| rms_norm_eps=1e-5, | |
| **kwargs | |
| ): | |
| super().__init__(**kwargs) | |
| self.vocab_size = vocab_size | |
| self.hidden_size = hidden_size | |
| self.num_hidden_layers = num_hidden_layers | |
| self.num_attention_heads = num_attention_heads | |
| self.intermediate_size = intermediate_size | |
| self.kv_head_split_layer = kv_head_split_layer | |
| self.num_kv_heads = num_kv_heads | |
| self.num_kv_heads_high = num_kv_heads_high | |
| self.tie_word_embeddings = tie_word_embeddings | |
| self.rms_norm_eps = rms_norm_eps | |