no self.image_token_posi and image_tokens

[hktk07]

In modeling_qwen2.py line 123 'HOLITOM_image_token_start_index = self.image_token_posi[0]' and 124 'HOLITOM_image_token_length = self.image_tokens[0]' have error of 'AttributeError: 'LlavaQwenModel' object has no attribute 'image_token_posi'' and 'AttributeError: 'LlavaQwenModel' object has no attribute 'image_tokens'. Did you mean: 'embed_tokens'?'

[cokeshao]

Sorry for the late response. It seems that AttributeError is caused by the different version of LLaVA. I wish you had fixed this problem.