Update modeling_phi2_clex.py

modeling_phi2_clex.py

@@ -59,7 +59,10 @@ logger = logging.get_logger(__name__)
 _CHECKPOINT_FOR_DOC = "microsoft/phi-2"
 _CONFIG_FOR_DOC = "CLEXPhiConfig"
 
-
+PHI_PRETRAINED_MODEL_ARCHIVE_LIST = [
+    "microsoft/phi-2",
+    # See all Phi models at https://huggingface.co/models?filter=phi
+]
 
 
 # Copied from transformers.models.llama.modeling_llama._get_unpad_data
@@ -373,10 +376,6 @@ class PhiAttention(nn.Module):
         # [batch_size, seq_length, num_heads, head_dim]
         query_states = torch.cat((query_rot, query_pass), dim=-1)
         key_states = torch.cat((key_rot, key_pass), dim=-1)
-        rotary_dim = int(self.partial_rotary_factor * self.head_dim)
-        if past_key_value is not None:
-            cache_kwargs = {"sin": sin, "cos": cos, "partial_rotation_size": rotary_dim}
-            key_states, value_states = past_key_value.update(key_states, value_states, self.layer_idx, cache_kwargs)
 
         key_states = repeat_kv(key_states, self.num_key_value_groups)
         value_states = repeat_kv(value_states, self.num_key_value_groups)