From 9f3e74813bea44411274e2f08c8ad1dcd1cc9e86 Mon Sep 17 00:00:00 2001
From: AllentDan <41138331+AllentDan@users.noreply.github.com>
Date: Mon, 15 Jul 2024 18:05:54 +0800
Subject: [PATCH] Fix internvl2-40b awq inference (#2023)

---
 lmdeploy/turbomind/deploy/source_model/internvl.py | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/lmdeploy/turbomind/deploy/source_model/internvl.py b/lmdeploy/turbomind/deploy/source_model/internvl.py
index 7f5a7c6ff..6b4dc2beb 100644
--- a/lmdeploy/turbomind/deploy/source_model/internvl.py
+++ b/lmdeploy/turbomind/deploy/source_model/internvl.py
@@ -94,6 +94,11 @@ class InternVLAwqReader(LlamaAwqReader):
     norm_weight_key = 'language_model.model.norm.weight'
     output_weight_key = 'language_model.lm_head.weight'
 
+    def __init__(self, new_params: dict, unused_params: dict, last_bin: bool,
+                 model_cfg: dict):
+        model_cfg = model_cfg.get('llm_config')
+        super().__init__(new_params, unused_params, last_bin, model_cfg)
+
 
 class InternVL2AwqReader(InternLM2AwqReader):
     """InternVLReader for InternLM2 model."""