Update modeling_hare.py

modeling_hare.py

@@ -1,9 +1,13 @@
 import json
+import logging
 from pathlib import Path
 
 import torch
-from transformers import AutoModel, AutoConfig, PreTrainedModel
-from transformers.modeling_outputs import BaseModelOutput
+from transformers import AutoModel, PreTrainedModel
+from transformers import ModernBertConfig
+
+for _logger_name in ["transformers.modeling_utils", "transformers.configuration_utils"]:
+    logging.getLogger(_logger_name).setLevel(logging.ERROR)
 
 from .configuration_hare import HareConfig
 from .birwkv7 import BiRWKV7Layer, init_from_attention
@@ -46,14 +50,20 @@ class HareModel(PreTrainedModel):
 
     def __init__(self, config):
         super().__init__(config)
-        base_config = AutoConfig.from_pretrained(
-            "answerdotai/ModernBERT-base",
+        base_config = ModernBertConfig(
             hidden_size=config.hidden_size,
             num_attention_heads=config.num_attention_heads,
             num_hidden_layers=config.num_hidden_layers,
             intermediate_size=config.intermediate_size,
             vocab_size=config.vocab_size,
             max_position_embeddings=config.max_position_embeddings,
+            pad_token_id=config.pad_token_id,
+            bos_token_id=config.bos_token_id,
+            eos_token_id=config.eos_token_id,
+            cls_token_id=getattr(config, 'cls_token_id', config.bos_token_id),
+            sep_token_id=getattr(config, 'sep_token_id', config.eos_token_id),
+            global_attn_every_n_layers=getattr(config, 'global_attn_every_n_layers', 3),
+            local_attention=getattr(config, 'local_attention', 128),
         )
         self.inner_model = AutoModel.from_config(base_config)