Hi everyone, I am following this tutorial Huggingface Knowledge Distillation and my process hangs when initializing DDP model this line
I add the following env variables NCCL_ASYNC_ERROR_HANDLING=1 NCCL_DEBUG=DEBUG TORCH_DISTRIBUTED_DEBUG=DETAIL for showing logs:
Here is the full log:
Traceback (most recent call last):
File "main.py", line 137, in <module>
main()
File "main.py", line 130, in main
distiller = Distiller(
File "/workspace/ProtonX-MLE/distiller.py", line 176, in __init__
self.student = DistributedDataParallel(
File "/root/miniconda3/envs/py38/lib/python3.8/site-packages/torch/nn/parallel/distributed.py", line 674, in __init__
_verify_param_shape_across_processes(self.process_group, parameters)
File "/root/miniconda3/envs/py38/lib/python3.8/site-packages/torch/distributed/utils.py", line 118, in _verify_param_shape_across_processes
return dist._verify_params_across_processes(process_group, tensors, logger)
RuntimeError: ProcessGroupWrapper: Monitored Barrier encountered error running collective: CollectiveFingerPrint(OpType=ALLGATHER, TensorShape=[1], TensorDtypes=Long, TensorDeviceTypes=TensorOptions(dtype=float (default), device=cuda, layout=Strided (default), requires_grad=false (default), pinned_memory=false (default), memory_format=(nullopt))). Error:
[../third_party/gloo/gloo/transport/tcp/pair.cc:598] Connection closed by peer [10.42.183.23]:11911
ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1) local_rank: 0 (pid: 1852) of binary: /root/miniconda3/envs/py38/bin/python
Traceback (most recent call last):
File "/root/miniconda3/envs/py38/lib/python3.8/runpy.py", line 194, in _run_module_as_main
return _run_code(code, main_globals, None,
File "/root/miniconda3/envs/py38/lib/python3.8/runpy.py", line 87, in _run_code
exec(code, run_globals)
File "/root/miniconda3/envs/py38/lib/python3.8/site-packages/torch/distributed/launch.py", line 196, in <module>
main()
File "/root/miniconda3/envs/py38/lib/python3.8/site-packages/torch/distributed/launch.py", line 192, in main
launch(args)
File "/root/miniconda3/envs/py38/lib/python3.8/site-packages/torch/distributed/launch.py", line 177, in launch
run(args)
File "/root/miniconda3/envs/py38/lib/python3.8/site-packages/torch/distributed/run.py", line 785, in run
elastic_launch(
File "/root/miniconda3/envs/py38/lib/python3.8/site-packages/torch/distributed/launcher/api.py", line 134, in __call__
return launch_agent(self._config, self._entrypoint, list(args))
File "/root/miniconda3/envs/py38/lib/python3.8/site-packages/torch/distributed/launcher/api.py", line 250, in launch_agent
raise ChildFailedError(
torch.distributed.elastic.multiprocessing.errors.ChildFailedError:
============================================================
main.py FAILED
------------------------------------------------------------
Failures:
<NO_OTHER_FAILURES>
------------------------------------------------------------
Root Cause (first observed failure):
[0]:
time : 2023-07-31_03:46:54
host : kb-dialogue-train-4-5cd5656dbd-qw7fl
rank : 0 (local_rank: 0)
exitcode : 1 (pid: 1852)
error_file: <N/A>
traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
============================================================
Can anyone help me out ?