About configs in 224_v2_llama2_video_stage_3.yaml

[tianguang2525]

1、Why in 224_v2_llama2_video_stage_3.yaml the config is llama_model: "meta-llama/Meta-Llama-3-8B-Instruct".While in stage_2, it is "meta-llama/Llama-2-7b-chat-hf".
2、Why the task is "image_text_pretrain" in stead of "fituning " in stage 3.
Thanks.