这是一篇记录多模态大模型和机器人环境交互(导航)实验过程的文章,本文为本系列的第一部分——环境搭建和数据集收集
暂记:vncserver启动:sudo vncserver :1 -geometry 1920x1080 -depth 24 -localhost no
Launch a controller
python -m llava.serve.controller --host 0.0.0.0 --port 10000
Launch a model worker
CUDA_VISIBLE_DEVICES=0 python -m llava.serve.model_worker --host 0.0.0.0 --controller http://localhost:10000 --port 40000 --worker http://localhost:40000 --model-path ~/LLaVA/llava-v1.5-7b-lora-ei --model-base ~/LLaVA/llava-v1.5-7b
Launch another model worker. (Using different port to avoid conflicts)
CUDA_VISIBLE_DEVICES=1 python -m llava.serve.model_worker --host 0.0.0.0 --controller http://localhost:10000 --port 30000 --worker http://localhost:30000 --model-path ~/LLaVA/llava-v1.5-7b