remove visual_inputs, fix qlora

examples/train_lora/llava1_5_lora_sft.yaml

@@ -1,6 +1,5 @@
 ### model
 model_name_or_path: llava-hf/llava-1.5-7b-hf
-visual_inputs: true
 
 ### method
 stage: sft

examples/train_lora/qwen2vl_lora_sft.yaml

@@ -1,6 +1,5 @@
 ### model
 model_name_or_path: Qwen/Qwen2-VL-7B-Instruct
-visual_inputs: true
 
 ### method
 stage: sft
@@ -9,7 +8,7 @@ finetuning_type: lora
 lora_target: all
 
 ### dataset
-dataset: mllm_demo
+dataset: mllm_demo,identity
 template: qwen2_vl
 cutoff_len: 1024
 max_samples: 1000