fix inter_size config

InternLM · Oct 23, 2024 · 4769ef8 · 4769ef8
1 parent 504093a
commit 4769ef8
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 1 deletion.
diff --git a/lmdeploy/turbomind/deploy/source_model/deepseek_vl.py b/lmdeploy/turbomind/deploy/source_model/deepseek_vl.py
@@ -47,7 +47,8 @@ def model_info(self):
                     'language_config'].get('model_type', None) == 'llama':
                 model_arg = model_arg['language_config']  # depseek-vl
             num_layer = model_arg['num_hidden_layers']
-            hidden_units = model_arg['hidden_size']
+            hidden_units = model_arg.get('hidden_size', 4096)
+            inter_size = model_arg.get('intermediate_size', 11008)
             norm_eps = model_arg.get('rms_norm_eps', 1e-06)
             attn_head_num = model_arg.get('num_attention_heads', 32)
             if 'num_key_value_heads' in model_arg:
@@ -71,6 +72,7 @@ def model_info(self):
                     head_num=attn_head_num,
                     kv_head_num=kv_head_num,
                     hidden_units=hidden_units,
+                    inter_size=inter_size,
                     rope_theta=rope_theta,
                     max_position_embeddings=max_position_embeddings,
                     use_dynamic_ntk=use_dynamic_ntk,

diff --git a/lmdeploy/turbomind/deploy/source_model/internvl.py b/lmdeploy/turbomind/deploy/source_model/internvl.py
@@ -62,6 +62,7 @@ def model_info(self):
             num_layer = model_arg['num_hidden_layers']
             norm_eps = model_arg['rms_norm_eps']
             hidden_units = model_arg['hidden_size']
+            inter_size = model_arg['intermediate_size']
             attn_head_num = model_arg['num_attention_heads']
             if 'num_key_value_heads' in model_arg:
                 kv_head_num = model_arg['num_key_value_heads']
@@ -82,6 +83,7 @@ def model_info(self):
         return dict(num_layer=num_layer,
                     norm_eps=norm_eps,
                     hidden_units=hidden_units,
+                    inter_size=inter_size,
                     head_num=attn_head_num,
                     kv_head_num=kv_head_num,
                     rope_theta=rope_theta,