Checkpoint: OverflowError: cannot serialize a string larger than 4GiB #1769
Labels
help wanted
Open to be worked on
question
Further information is requested
won't fix
This will not be worked on
🐛 Bug
Model checkpointing fails with the error: OverflowError: cannot serialize a string larger than 4GiB and halts training
conda
,pip
, source): condaAdditional context
This is a known Python issue pytorch/pytorch#12085
Possible fix, set the protocol correctly
The text was updated successfully, but these errors were encountered: