ConnectionRefusedError: [Errno 111] Connection refused

[DouPiChen]

❓ The question

When I run torchrun --nproc_per_node=4 scripts/train.py configs/official/OLMo-1B.yaml, I get the ConnectionRefusedError as follows

I have changed url in OLMo-1B.yaml as

[2015aroras]

It's hard to tell what the error is from jus the call stack, which is almost all inside python's urllib3 error. Based on the last INFO line in the first image, the error appears to be in the Weights & Biases SDK. If you do not need Weights & Biases to keep track of run metrics, you can set --wandb to null. That is, run torchrun --nproc_per_node=4 scripts/train.py configs/official/OLMo-1B.yaml --wandb=null.

[dumitrac]

I apologize for our delay in response. In order to help surface current, unresolved issues, we are closing tickets prior to February 29. Please reopen your ticket if you are continuing to experience this issue. Thank you!