Commit 2f3a2d68 authored by Jared Casper's avatar Jared Casper
Browse files

Merge branch 'extra_assertion' into 'main'

Make sure pipeline-model-parallel size is greater than 2 for interleaved schedule

See merge request ADLR/megatron-lm!253
parents a6e00d97 182841f7
Loading
Loading
Loading
Loading
+3 −0
Original line number Diff line number Diff line
@@ -116,6 +116,9 @@ def parse_args(extra_args_provider=None, defaults={},
                args.global_batch_size), flush=True)
    assert args.global_batch_size > 0
    if args.num_layers_per_virtual_pipeline_stage is not None:
        assert args.pipeline_model_parallel_size > 2, \
            'pipeline-model-parallel size should be greater than 2 with ' \
            'interleaved schedule'
        assert args.num_layers % args.num_layers_per_virtual_pipeline_stage == 0, \
            'number of layers is not divisible by number of layers per virtual ' \
            'pipeline stage'