use seen_tokens in DynamicCache

modeling_deepseek.py

@@ -1724,7 +1724,7 @@ class DeepseekV2ForCausalLM(DeepseekV2PreTrainedModel):
         if past_key_values is not None:
             if isinstance(past_key_values, Cache):
                 cache_length = past_key_values.get_seq_length()
-                past_length = past_key_values.cache_position
+                past_length = past_key_values.seen_tokens
                 max_cache_length = past_key_values.get_max_cache_shape()
             else:
                 cache_length = past_length = past_key_values[0][0].shape[2]