关于您的论文复现

Smallwhite999 commented 4 years ago

作者您好，我是按照 README.md 上面的配置的，训练和推理也是按照那个里面的命令执行的，但是还是报错，是除了按照 README.md 里面的配资之外还要按照哪个配置啊？劳烦解答一下，谢谢！

Scalsol commented 4 years ago

请讲下你的环境 (python版本, nvcc, gcc等), 然后讲一下是运行什么命令导致报错了, 谢谢!

Smallwhite999 commented 4 years ago

我安装的时候都是按照下面这个安装的：

训练的命令是按照这个操作的：这些配置完成之后报错是这样的： python -m torch.distributed.launch \

--nproc_per_node=2 \ tools/train_net.py \ --master_port=$((RANDOM+10000)) \ --config-file configs/MEGA/vid_R_101_c4_MEGA_1x.yaml \ --motion-specific

Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed.

/home/wang/anaconda3/envs/cvbaseline/bin/python: can't open file 'tools/train_net.py': [Errno 2] No such file or directory Traceback (most recent call last): File "/home/wang/anaconda3/envs/cvbaseline/lib/python3.8/runpy.py", line 194, in _run_module_as_main return _run_code(code, main_globals, None, File "/home/wang/anaconda3/envs/cvbaseline/lib/python3.8/runpy.py", line 87, in _run_code exec(code, run_globals) File "/home/wang/anaconda3/envs/cvbaseline/lib/python3.8/site-packages/torch/distributed/launch.py", line 263, in main() File "/home/wang/anaconda3/envs/cvbaseline/lib/python3.8/site-packages/torch/distributed/launch.py", line 258, in main raise subprocess.CalledProcessError(returncode=process.returncode, subprocess.CalledProcessError: Command '['/home/****/anaconda3/envs/cvbaseline/bin/python', '-u', 'tools/train_net.py', '--local_rank=1', '--master_port=16269', '--config-file', 'configs/MEGA/vid_R_101_c4_MEGA_1x.yaml', '--motion-specific']' returned non-zero exit status 2. /home/wang/anaconda3/envs/cvbaseline/bin/python: can't open file 'tools/train_net.py': [Errno 2] No such file or directory 并且我跳过这个命令“--motion-specific”的时候报错是这样的： python -m torch.distributed.launch \ --nproc_per_node=2 \ tools/train_net.py \ --master_port=$((RANDOM + 10000)) \ --config-file configs/MEGA/vid_R_101_C4_MEGA_1x.yaml \ OUTPUT_DIR training_dir/MEGA_R_101_1x

Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed.

/home/wang/anaconda3/envs/cvbaseline/bin/python: can't open file 'tools/train_net.py': [Errno 2] No such file or directory Traceback (most recent call last): File "/home/wang/anaconda3/envs/cvbaseline/lib/python3.8/runpy.py", line 194, in _run_module_as_main return _run_code(code, main_globals, None, /home/wang/anaconda3/envs/cvbaseline/bin/python: can't open file 'tools/train_net.py': [Errno 2] No such file or directory File "/home/wang/anaconda3/envs/cvbaseline/lib/python3.8/runpy.py", line 87, in _run_code exec(code, run_globals) File "/home/wang/anaconda3/envs/cvbaseline/lib/python3.8/site-packages/torch/distributed/launch.py", line 263, in main() File "/home/wang/anaconda3/envs/cvbaseline/lib/python3.8/site-packages/torch/distributed/launch.py", line 258, in main raise subprocess.CalledProcessError(returncode=process.returncode, subprocess.CalledProcessError: Command '['/home/wangdicong19b/anaconda3/envs/cvbaseline/bin/python', '-u', 'tools/train_net.py', '--local_rank=1', '--master_port=37600', '--config-file', 'configs/MEGA/vid_R_101_C4_MEGA_1x.yaml', 'OUTPUT_DIR', 'training_dir/MEGA_R_101_1x']' returned non-zero exit status 2.

在2020年7月23日 17:21，Yihong Chennotifications@github.com 写道：

请讲下你的环境 (python版本, nvcc, gcc等), 然后讲一下是运行什么命令导致报错了, 谢谢!

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

Scalsol commented 4 years ago

你运行命令的时候切换到mega.pytorch这个目录下了吗?

Smallwhite999 commented 4 years ago

我再试一下吧，谢谢您的解答！在2020年7月24日 10:33，Yihong Chennotifications@github.com 写道：

你运行命令的时候切换到mega.pytorch这个目录下了吗?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

launchauto commented 4 years ago

我再试一下吧，谢谢您的解答！在2020年7月24日 10:33，Yihong Chennotifications@github.com 写道：你运行命令的时候切换到mega.pytorch这个目录下了吗? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

请问你复现成功了么

Smallwhite999 commented 4 years ago

并没有在2020年8月26日 17:20，LauncHnotifications@github.com 写道：

我再试一下吧，谢谢您的解答！在2020年7月24日 10:33，Yihong Chennotifications@github.com 写道：你运行命令的时候切换到mega.pytorch这个目录下了吗? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

请问你复现成功了么

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

fakechen commented 4 years ago

并没有在2020年8月26日 17:20，LauncHnotifications@github.com 写道：我再试一下吧，谢谢您的解答！在2020年7月24日 10:33，Yihong Chennotifications@github.com 写道：你运行命令的时候切换到mega.pytorch这个目录下了吗? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe. 请问你复现成功了么 — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

请问复现成功了吗

fakechen commented 4 years ago

在collate_batch.py中 transposed_batch[0][0]的数据结构tensor 为什么用字典操作啊