What is the difference between it and lmms-lab/llava-onevision-qwen2-72b-ov?
#6
by
wyccccc
- opened
Does lmms-lab/llava-onevision-qwen2-72b-ov-sft exhibit a superior ability to follow instructions?
When I used lmms-lab/llava-onevision-qwen2-72b-ov directly, I noticed that it generated outputs that seemed almost random, with very little adherence to the given instructions. I'm unsure whether this is due to incorrect usage of the model version or some other factor. Could you provide insights into what might be causing this issue and how I can improve its instruction-following performance?