Nvidia Megatron Bridge Perf Sequence Packing

Sequence Packing Skill

For stable background and recommendation level, see:

• @docs/training/packed-sequences.md
• @skills/nvidia-megatron-bridge-perf-sequence-packing/card.yaml

Enablement

Offline packed SFT for LLM finetuning:

from megatron.bridge.data.datasets.packed_sequence import PackedSequenceSpecs

cfg.train.micro_batch_size = 1
cfg.dataset.seq_length = 4096
cfg.model.seq_length = 4096
cfg.dataset.dataset_kwargs = {"pad_to_max_length": True}
cfg.dataset.packed_sequence_specs = PackedSequenceSpecs(
    packed_sequence_size=4096,
    pad_seq_to_mult=1,
)

If CP is enabled:

cfg.model.context_parallel_size = 2
cfg.model.calculate_per_token_loss = True
cfg.ddp.average_in_collective = False
cfg.dataset.packed_sequence_specs.pad_seq_to_mult = cfg.model.context_parallel_size * 2

# If sequence_parallel is also enabled, use lcm(2*CP, CP*TP):
# import math
# cfg.dataset.packed_sequence_specs.pad_seq_to_mult = math.lcm(2 * CP, CP * TP)
# See src/megatron/bridge/training/vlm_step.py for reference logic.

If CUDA graphs are enabled for this packed path:

cfg.dataset.packed_sequence_specs.pad_cu_seqlens = True
cfg.dataset.dataset_kwargs["pad_to_max_length"] = True

Note: pad_cu_seqlens = True also requires a metadata JSON file alongside the packed dataset (asserted in src/megatron/bridge/data/datasets/sft.py). Custom packed datasets that omit the metadata file will hit an assertion at dataset initialization.

In-batch packing for VLM finetuning:

cfg.dataset.pack_sequences_in_batch = True
cfg.train.micro_batch_size = 2

Long-context baseline:

cfg.model.seq_length = 16384
cfg.dataset.seq_length = 16384
cfg.model.context_parallel_size = 2

Code Anchors

LLM packed SFT config surface:

if packed_sequence:
    dataset_kwargs = {"pad_to_max_length": True}
    packed_sequence_specs = PackedSequenceSpecs(packed_sequence_size=seq_length, pad_seq_to_mult=pad_seq_to_mult)
else:
    dataset_kwargs = {}
    packed_sequence_specs = None

Bridge validation:

if self.model.context_parallel_size > 1:
    assert self.model.seq_length % (self.model.context_parallel_size * 2) == 0, ...
    if isinstance(self.dataset, FinetuningDatasetConfig):
        assert self.model.calculate_per_token_loss, ...
        assert not self.ddp.average_in_collective, ...
...
if ... packed_sequence_size > 0 and self.train.micro_batch_size > 1:
    raise ValueError(...)
...
if getattr(self.dataset, "pack_sequences_in_batch", False) and self.train.micro_batch_size == 1:
    raise ValueError(...)

VLM in-batch runtime:

if enable_packing:
    ...
    ) = pack_batch_sequences(
        ...
        pad_token_id=0,
        pad_to_multiple_of=cp_size * 2 if cp_size > 1 else 1,
    )

Packed THD runtime constraint:

if cu_seqlens.dim() > 1 and cu_seqlens.size(0) != 1:
    raise ValueError("Packed THD batches expect micro-batch size 1 for context-parallel slicing (THD layout)")

Pitfalls

1. Offline packed SFT and VLM in-batch packing are different features with opposite micro-batch rules.
2. When CP is enabled, packed sequence lengths must respect 2 * context_parallel_size divisibility.
3. For finetuning with CP, calculate_per_token_loss=True and ddp.average_in_collective=False are required.
4. pad_cu_seqlens=True also requires pad_to_max_length=True.
5. Packing support is model-family-specific. Qwen3-Next, GLM-4.5, and Qwen3.5-VL contain explicit opt-outs in different paths.
6. MTP finetuning is documented as incompatible with packed sequences.

Verification

Use the checked-in unit coverage:

uv run python -m pytest tests/unit_tests/training/utils/test_packed_seq_utils.py -v && \
uv run python -m pytest tests/unit_tests/training/test_config.py -k "packed_sequence or pack_sequences_in_batch or context_parallel_seq_length_divisibility or context_parallel_finetuning_validations" -v && \
uv run python -m pytest tests/unit_tests/training/test_vlm_step.py -k "enable_packing" -v

Success criteria:

• first command reports 8 passed
• second command reports 14 passed
• third command reports 2 passed