Open XiaoyiHan6 opened 5 months ago
Dear Zhang, 我最近比较忙,才看到邮箱,如果你想知道work_dir是如何定义的,你必须找到test_relation_full.sh中的配置.py的文件 ”# sh scripts/test/test_relation_full.sh PARTITION=priority JOB_NAME=psg PORT=${PORT:-$((29500 + $RANDOM % 29))} GPUS_PER_NODE=${GPUS_PER_NODE:-1} CPUS_PER_TASK=${CPUS_PER_TASK:-5}
PYTHONPATH="/mnt/lustre/jkyang/CVPR23/openpvsg":$PYTHONPATH \ srun -p ${PARTITION} \ --job-name=${JOB_NAME} \ --gres=gpu:${GPUS_PER_NODE} \ --ntasks-per-node=${GPUS_PER_NODE} \ --cpus-per-task=${CPUS_PER_TASK} \ --kill-on-bad-exit=1 \ python tools/rel_test_full.py --launcher="slurm" ${PY_ARGS}“ 正如这段代码所示它的配置文件为” tools/rel_test_full.py“找到这个rel_test_full.py中的args就可以知道配置文件的定义了。 Best Wishes, Xiaoyi Han
在 2024-10-10 19:42:41,"Zhen Pang" @.***> 写道:
你好,请问这里的work_dir参数是如何定义的呢
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>
Dear Zhang, 我最近比较忙,才看到邮箱,如果你想知道work_dir是如何定义的,你必须找到test_relation_full.sh中的配置.py的文件 ”# sh scripts/test/test_relation_full.sh PARTITION=priority JOB_NAME=psg PORT=${PORT:-$((29500 + $RANDOM % 29))} GPUS_PER_NODE=${GPUS_PER_NODE:-1} CPUS_PER_TASK=${CPUS_PER_TASK:-5} PYTHONPATH="/mnt/lustre/jkyang/CVPR23/openpvsg":$PYTHONPATH \ srun -p ${PARTITION} \ --job-name=${JOB_NAME} \ --gres=gpu:${GPUS_PER_NODE} \ --ntasks-per-node=${GPUS_PER_NODE} \ --cpus-per-task=${CPUS_PER_TASK} \ --kill-on-bad-exit=1 \ python tools/rel_test_full.py --launcher="slurm" ${PY_ARGS}“ 正如这段代码所示它的配置文件为” tools/rel_test_full.py“找到这个rel_test_full.py中的args就可以知道配置文件的定义了。 Best Wishes, Xiaoyi Han 在 2024-10-10 19:42:41,"Zhen Pang" @.> 写道: 你好,请问这里的work_dir参数是如何定义的呢 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.> 尊敬的 Zhang, 我最近比较忙,才看到邮箱,如果你想知道work_dir是如何定义的,你必须找到test_relation_full.sh中的配置.py的文件 “# sh scripts/test/test_relation_full.sh PARTITION=优先级 JOB_NAME=psg 端口=${端口:-$((29500 + $RANDOM % 29))}} GPUS_PER_NODE=${GPUS_PER_NODE:-1} CPUS_PER_TASK=${CPUS_PER_TASK:-5}
PYTHONPATH=“/mnt/lustre/jkyang/CVPR23/openpvsg”:$PYTHONPATH \ srun -p ${分区} \ --job-name=${JOB_NAME} \ --gres=gpu:${GPUS_PER_NODE} \ --ntasks 每节点=${GPUS_PER_NODE} \ --cpus-per-task=${CPUS_PER_TASK} \ --kill-on-bad-exit=1 \ python tools/rel_test_full.py --launcher="slurm" ${PY_ARGS}“ 正如这段代码所示它的配置文件为” tools/rel_test_full.py“找到这个rel_test_full.py中的args就可以知道配置文件的定义了。 愿你安好 韩小义
在 2024-10-10 19:42:41,"Zhen Pang" @.***> 写道:
你好,请问这里的work_dir参数是如何定义的呢
— 直接回复此电子邮件,在 GitHub 上查看,或取消订阅。 您收到此消息是因为您编写了该线程。消息 ID:@.***>
Thank you very much,I have solved it.
Hello Dr. Yang Your original code test_relation_full.sh
The code is incorrect, because your tools/rel_test_full.py doesn't even have an argument for args.launcher. And This is an error message.
your original code tools/rel_test_full.py
So, when we execute test_relation_full.sh, we should delete this argument
--launcher="slurm"
I wrote this issue to tell anyone who executes this code to be aware of this bug.
Sincerely, XY.Han