autohandai/community-skills - Page 2

nvidia-megatron-bridge-adding-model-support

Guide for adding support for new LLM or VLM models in Megatron-Bridge. Covers bridge, provider, recipe, tests, docs, and examples.

nvidia-megatron-bridge-build-and-dependency

Dev environment setup for Megatron Bridge — container-based development, uv package management, lockfile regeneration, adding dependencies, Slurm container usage, and common build pitfalls.

nvidia-megatron-bridge-bump-dependency

Bump a pinned dependency (TransformerEngine, Megatron-LM, NRX, etc.), regenerate the lockfile, open a PR, and drive it to green by attaching a watchdog to the "CICD NeMo" workflow and quarantining failing functional tests as flaky until the run is green.

nvidia-megatron-bridge-cicd

CI/CD reference for Megatron Bridge — pipeline structure, commit and PR workflow, CI failure investigation, and common failure patterns.

nvidia-megatron-bridge-linting-and-formatting

Code style and quality rules for Megatron Bridge — ruff configuration, naming conventions, type hints, mypy rules, docstrings, copyright headers, logging, and the code review checklist.

nvidia-megatron-bridge-mlm-bridge-training

Run Megatron-LM (MLM) and Megatron Bridge training with mock or real data. Covers correlation testing, available recipes, and multi-GPU examples.

nvidia-megatron-bridge-multi-node-slurm

Convert single-node scripts to multi-node Slurm sbatch jobs and debug common multi-node failures. Covers srun-native vs uv run torch.distributed approaches, container setup, NCCL timeouts, OOM sizing for MoE models, and interactive allocation.

nvidia-megatron-bridge-nemo-rl-e2e-testing

External NeMo-RL end-to-end validation workflow for Megatron-Bridge model/provider changes, including downstream compatibility checks, external RL lifecycle behavior, Megatron policy setup, HF import/export, checkpoint/resume, non-colocated vLLM refit, delta…

nvidia-megatron-bridge-parity-testing

Structured framework for verifying numerical parity of HF-to-MCore weight conversions. References existing tools and the add-model-support skill.

nvidia-megatron-bridge-perf-activation-recompute

Validate and use selective and full activation recompute in Megatron Bridge to reduce GPU memory usage at the cost of extra compute.

nvidia-megatron-bridge-perf-cpu-offloading

Validate and use CPU offloading in Megatron Bridge, including layer-level activation offloading and fractional optimizer state offloading with HybridDeviceOptimizer.

nvidia-megatron-bridge-perf-cuda-graphs

Validate and use CUDA graph capture in Megatron Bridge, including local full-iteration graphs and Transformer Engine scoped graphs for attention, MLP, and MoE modules.

nvidia-megatron-bridge-perf-expert-parallel-overlap

Validate and use MoE expert-parallel communication overlap in Megatron-Bridge, including overlap_moe_expert_parallel_comm, delay_wgrad_compute, and flex dispatcher backends such as DeepEP and HybridEP.

nvidia-megatron-bridge-perf-hierarchical-context-parallel

Operational guide for enabling hierarchical context parallelism in Megatron-Bridge, including config knobs, code anchors, pitfalls, and verification.

nvidia-megatron-bridge-perf-megatron-fsdp

Operational guide for enabling Megatron FSDP in Megatron-Bridge, including config knobs, code anchors, pitfalls, and verification.

nvidia-megatron-bridge-perf-memory-tuning

Techniques for reducing peak GPU memory in Megatron Bridge — expandable segments, parallelism resizing, activation recompute, CPU offloading constraints, and common OOM fixes.

nvidia-megatron-bridge-perf-moe-comm-overlap

MoE expert-parallel communication overlap in Megatron Bridge. Covers dispatch/combine overlap, flex dispatcher backends, and expert wgrad scheduling.

nvidia-megatron-bridge-perf-moe-dispatcher-selection

Choose the right MoE token dispatcher (`alltoall`, DeepEP, or HybridEP) for the hardware, EP degree, and optimization stage. Summarizes patterns from DSV3, Qwen3, Qwen3-Next, and VLM bring-up work.

nvidia-megatron-bridge-perf-moe-hardware-configs

Representative MoE training playbooks by hardware platform and model family. Summarizes rounded throughput bands, parallelism patterns, and common tuning stacks.

nvidia-megatron-bridge-perf-moe-long-context

Long-context MoE training guidance for Megatron Bridge. Covers CP sizing, selective recompute, dispatcher choices, and practical patterns from DSV3, Qwen3, and Qwen3-Next long-context experiments.

nvidia-megatron-bridge-perf-moe-optimization-workflow

Systematic workflow for MoE training optimization in Megatron Bridge, based on the Megatron-Core MoE paper. Covers the Three Walls framework, parallel folding, recompute strategy, dispatcher choice, and CUDA-graph bring-up.

nvidia-megatron-bridge-perf-moe-vlm-training

Practical guidance for training MoE VLMs in Megatron Bridge. Compares FSDP and 3D-parallel approaches, using rounded lessons from Qwen3-VL, Qwen3-Next, and other multimodal experiments.

nvidia-megatron-bridge-perf-parallelism-strategies

Operational guide for choosing and combining parallelism strategies in Megatron Bridge, including sizing rules, hardware topology mapping, and combined parallelism configuration.

nvidia-megatron-bridge-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.

nvidia-megatron-bridge-perf-tp-dp-comm-overlap

Operational guide for enabling TP, DP, and PP communication overlap in Megatron-Bridge, including config knobs, code anchors, pitfalls, and verification.

nvidia-megatron-bridge-recipe-recommender

Recommend and customize Megatron Bridge recipes for a user's model, GPU count, and training goal. Indexes library recipes (pretrain/SFT/PEFT) and performance recipes.

nvidia-megatron-bridge-resiliency

Resiliency features in Megatron Bridge including fault tolerance, straggler detection, in-process restart, preemption, and re-run state machine.

nvidia-megatron-bridge-testing

Testing reference for Megatron Bridge — unit and functional test layout, tier semantics (L0/L1/L2/flaky), script conventions, running tests locally, adding/moving/disabling tests, and pytest conventions.

nvidia-megatron-bridge-verl-e2e-testing

External verl end-to-end validation workflow for Megatron-Bridge model/provider changes. Covers running a small verl Megatron backend job from a Bridge checkout, choosing LoRA/DDP plus optional save/resume and parallelism variants, setting PYTHONPATH so verl…

nvidia-megatron-core-build-and-dependency

Container-based dev environment setup and dependency management for Megatron-LM. Covers acquiring and launching the CI container, uv package management, and updating uv.lock.

nvidia-megatron-core-bump-base-image

Bump the NVIDIA PyTorch base image using the nvcr.io/nvidia/pytorch YY.MM py3 tag used by Megatron-LM CI. Covers the two pin sites (GitHub CI in `docker/.ngc_version.dev` and GitLab CI in `.gitlab/stages/01.build.yml`), the post-bump CI loop (re-run…

nvidia-megatron-core-cicd

CI/CD reference for Megatron-LM. Covers CI pipeline structure, PR scope labels, triggering internal GitLab CI, and CI failure investigation.

nvidia-megatron-core-create-issue

Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.

nvidia-megatron-core-linting-and-formatting

Linting and formatting for Megatron-LM. Covers running autoformat.sh, tools (ruff, black, isort, pylint, mypy), and code style rules.

nvidia-megatron-core-nightly-sync

Domain knowledge for the nightly main-to-dev sync workflow. Covers merge strategy, CI architecture, failure investigation, and known issues.

nvidia-megatron-core-onboard-gb200-1node-tests

Onboard 1-node GitHub MR functional tests for GB200 from existing mr-scoped 2-node tests.

nvidia-megatron-core-respond-to-issue

Research and draft a response to a GitHub issue or question from an external contributor.

nvidia-megatron-core-run-on-slurm

How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions,…

nvidia-megatron-core-split-pr

Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.

nvidia-megatron-core-testing

Test system for Megatron-LM. Covers test layout, recipe YAML structure, adding and running unit and functional tests, golden values, marker filters, and CI parity.