nemo-mbridge-perf-sequence-packing
Validate and use packed sequences and long-context training in Megatron-Bridge, distinguishing offline packed SFT for LLMs from in-batch packing for VLMs, and applying the right CP constraints.
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-10
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- nemo-mbridge-perf-sequence-packing
- license
- Apache-2.0
- description
- Validate and use packed sequences and long-context training in Megatron-Bridge, distinguishing offline packed SFT for LLMs from in-batch packing for VLMs, and applying the right CP constraints.
- when_to_use
- Enabling sequence packing or long-context SFT, or investigating a commit that broke sequence packing or changed packing behavior; 'packed sequences', 'sequence packing', 'PackedSequenceSpecs', 'pack_sequences_in_batch', 'CP with packing'.
02 资源文件 5
- 参考文档BENCHMARK.md4.3 KB
- 参考文档card.yaml4.1 KB
- 参考文档evals/evals.json1.8 KB
- 参考文档skill-card.md4.0 KB
- 其他skill.oms.sig4.7 KB
03 SKILL.md 正文
Sequence Packing Skill
For stable background and recommendation level, see:
- @docs/training/packed-sequences.md
- @skills/nemo-mbridge-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
- Offline packed SFT and VLM in-batch packing are different features with opposite micro-batch rules.
- When CP is enabled, packed sequence lengths must respect
2 * context_parallel_sizedivisibility. - For finetuning with CP,
calculate_per_token_loss=Trueandddp.average_in_collective=Falseare required. pad_cu_seqlens=Truealso requirespad_to_max_length=True.- Packing support is model-family-specific.
Qwen3-Next,GLM-4.5, andQwen3.5-VLcontain explicit opt-outs in different paths. - 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
04 同仓库更多 Skills
vercel-sandbox
Run agent-browser + Chrome inside Vercel Sandbox microVMs for browser automation from any Vercel-deployed app. Use when the user needs browser automation in a Vercel app (Next.js, SvelteKit, Nuxt, Remix, Astro, etc.), wants to run headless Chrome without binary size limits, needs persistent browser sessions across commands, or wants ephemeral isolated browser environments. Triggers include "Vercel Sandbox browser", "microVM Chrome", "agent-browser in sandbox", "browser automation on Vercel", or any task requiring Chrome in a Vercel Sandbox.
slack
Interact with Slack workspaces using browser automation. Use when the user needs to check unread channels, navigate Slack, send messages, extract data, find information, search conversations, or automate any Slack task. Triggers include "check my Slack", "what channels have unreads", "send a message to", "search Slack for", "extract from Slack", "find who said", or any task requiring programmatic Slack interaction.
electron
Automate Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify, etc.) using agent-browser via Chrome DevTools Protocol. Use when the user needs to interact with an Electron app, automate a desktop app, connect to a running app, control a native app, or test an Electron application. Triggers include "automate Slack app", "control VS Code", "interact with Discord app", "test this Electron app", "connect to desktop app", or any task requiring automation of a native Electron application.
dogfood
Systematically explore and test a web application to find bugs, UX issues, and other problems. Use when asked to "dogfood", "QA", "exploratory test", "find issues", "bug hunt", "test this app/site/platform", or review the quality of a web application. Produces a structured report with full reproduction evidence -- step-by-step screenshots, repro videos, and detailed repro steps for every issue -- so findings can be handed directly to the responsible teams.
agent-browser
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.
shadcn
Manages shadcn components and projects — adding, searching, fixing, debugging, styling, and composing UI, including chat interfaces. Provides project context, component docs, and usage examples. Applies when working with shadcn/ui, component registries, presets, --preset codes, or any project with a components.json file. Also triggers for "shadcn init", "create an app with --preset", "switch to --preset", and for building a design system or applying a DESIGN.md or brand spec to shadcn/ui.