Commit 217f54b3 authored by Mohammad Shoeybi's avatar Mohammad Shoeybi
Browse files

Merge branch 'finetune_assert' into 'main' 

Add temporary assert to finetuning until it can be fixed.

See merge request ADLR/megatron-lm!278
parents 42c1cf42 113c6362

tasks/finetune_utils.py

+3 −1
Original line number Diff line number Diff line
@@ -19,7 +19,7 @@ from functools import partial 


import torch



from megatron import get_args

from megatron import get_args, get_num_microbatches

from megatron import print_rank_0

from megatron import get_timers

from megatron import mpu
@@ -154,6 +154,8 @@ def _train(model, optimizer, lr_scheduler, forward_step, 
    args = get_args()

    timers = get_timers()



    assert get_num_microbatches() == 1, "finetuning with gradient accumulation doesn't currently work"



    # Turn on training mode which enables dropout.

    for m in model:

        m.train()