Using the DDP example here DDP-based Scaling of Gaudi on PyTorch — Gaudi Documentation 1.6.0 documentation Tried to replace spawn with fork using the following changes resulted in an error: import os import sys import tempfile import torch import torch.distributed as dist import torch.nn as n…

Using DDP with fork fails

Sayantan_S August 18, 2022, 7:57pm 4

I tried it on 1.6 and it seems to work. Can you please try it on 1.6 and see if it helps.

Topic		Replies	Views
Trainer killed/Segfault PyTorch	6	603	September 1, 2023
Model.to device faile: "RuntimeError: synStatus=8 [Device not found] Device acquire failed." Training models , pytorch	3	582	March 13, 2024
Hccl failure to be connected on two nodes with a simple script PyTorch	1	66	January 2, 2025
Can I train DistilBERT with Gaudi? Training	3	802	January 12, 2022
RuntimeError: [Rank:0] FATAL ERROR :: MODULE:PT_BRIDGE Exception in Lowering thread PyTorch	3	200	March 23, 2025