The inference process is stuck. about multi-task-transformer

[prismformore]

Hi, thank you for your interests. I have not heard about this problem. Can you locate at which line does it stop? It seems that the pre-trained model is not successfullly loaded because it says "infer at iteration 0." Hanrong Get Outlook for Android<https://aka.ms/AAb9ysg>

…

________________________________ From: Jiangpengtao ***@***.***> Sent: Wednesday, May 24, 2023 7:03:15 PM To: prismformore/Multi-Task-Transformer ***@***.***> Cc: Subscribed ***@***.***> Subject: [prismformore/Multi-Task-Transformer] The inference process is stuck. (Issue #20) When I use the pre-trained model on NYUD-v2, the inference process is stuck. Here is my command CUDA_VISIBLE_DEVICES=0 python -m torch.distributed.launch \ --nproc_per_node=1 \ --master_port=$((RANDOM%1000+12000)) \ main.py \ --config_exp './configs/nyud/nyud_vitLp16_taskprompter.yml' \ --run_mode infer \ --trained_model TaskPrompter_nyud_vitLp16.pth.tar The log is as follows: Number of dataset images: 654 Use checkpoint TaskPrompter_nyud_vitLp16.pth.tar Infer at iteration 0 0%| — Reply to this email directly, view it on GitHub<#20>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADLPY6STEWMG3S4SVPHEYVTXH24WHANCNFSM6AAAAAAYOFLO7U>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[PengtaoJiang]

Hi, at L448 in taskprompter.py, cur_task_fea = self.fea_decode_spa[il][task](cur_task_fea)

[PengtaoJiang]

Hello, hanrong,
when I use the inference.py file, the inference process is running successfully. Here is my running command.

CUDA_VISIBLE_DEVICES=0 python inference.py \
    --config_path=./configs/nyud/nyud_vitLp16_taskprompter.yml \
    --image_path=./dataset/NYUDv2/images/0001.png \
    --ckp_path=./TaskPrompter_nyud_vitLp16.pth.tar \
    --save_dir=../TaskPrompter_nyud_vitLp16/results/

[prismformore]

Hi, at L448 in taskprompter.py, cur_task_fea = self.fea_decode_spa[il][task](cur_task_fea)

That is quite interesting. Does it result from the ddp? Are you using the recommended environment?

[PengtaoJiang]

Hello, I guess the problem occurs when I use DDP with one GPU.

[prismformore]

@PengtaoJiang This should not be a problem as I always use it in this way. If you use the recommended environment does this still happen?

[PengtaoJiang]

@prismformore,
hello, hanrong, I successfully run the evaluation code. I guess it is my MACHINE problem.