LLaVA-Next-3D/data_precessing/llm_analyze.sh

#!/bin/bash

export python3WARNINGS=ignore
export TOKENIZERS_PARALLELISM=false

srun -p mozi-S1 --gres=gpu:8 -N1 -n1 --cpus-per-task=12 python data_precessing/ray_analyze.py

# python llava/eval/eval_scanrefer.py --input-file results/scanrefer/prompt-llavanext-qwen-3d-uniform-32gpu-double-vg-clip.jsonl
# sh scripts/3d/eval/eval_scanrefer.sh llavanext-qwen-3d-uniform/checkpoint-6500 uniform 32