TR

training-llms-megatron

Framework for large-scale distributed LLM training with high hardware utilization on NVIDIA GPUs.

Install

mkdir -p .claude/skills/training-llms-megatron && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/7702" && unzip -o skill.zip -d .claude/skills/training-llms-megatron && rm skill.zip

Installs to .claude/skills/training-llms-megatron

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies. Use when training models >1B parameters, need maximum GPU efficiency (47% MFU on H100), or require tensor/pipeline/sequence/context/expert parallelism. Production-ready framework used for Nemotron, LLaMA, DeepSeek.
329 chars✓ has a “when” triggerlonger than Claude Code's old 250-char listing cap (fine on current versions)
Advanced

Key capabilities

  • Orchestrates distributed 3D parallelism (tensor, pipeline, sequence)
  • Optimizes GPU memory usage for models >1B parameters
  • Launches multi-node training jobs
  • Monitors H100 efficiency metrics

How it works

Manages communication primitives between GPU groups to shard parameters and gradients across a distributed cluster.

Inputs & outputs

You give it
Training config file and model architecture
You get back
Distributed training environment state

When to use training-llms-megatron

  • Configure model parallelism
  • Launch distributed training
  • Monitor training performance

About this skill

Megatron-Core - Large-Scale LLM Training

Quick start

Megatron-Core trains LLMs from 2B to 462B parameters with up to 47% Model FLOP Utilization on H100 GPUs through advanced parallelism strategies.

Installation:

# Docker (recommended)
docker run --gpus all -it --rm nvcr.io/nvidia/pytorch:25.04-py3

# Or pip
pip install megatron-core

Simple distributed training:

# Train with 2 GPUs using data parallelism
torchrun --nproc_per_node=2 examples/run_simple_mcore_train_loop.py

# Or LLaMA-3 8B training
./examples/llama/train_llama3_8b_fp8.sh

Common workflows

Workflow 1: Train LLaMA-style model with 3D parallelism

Copy this checklist:

LLaMA Training Setup:
- [ ] Step 1: Choose parallelism configuration
- [ ] Step 2: Configure training hyperparameters
- [ ] Step 3: Launch distributed training
- [ ] Step 4: Monitor performance metrics

Step 1: Choose parallelism configuration

Model size determines parallelism strategy:

Model SizeGPUsTensor ParallelPipeline ParallelData ParallelContext Parallel
7B81181
13B82141
70B644441
405B1288822

Step 2: Configure training hyperparameters

#!/bin/bash
# train_llama_70b.sh

GPUS_PER_NODE=8
NNODES=8  # 64 GPUs total
TP=4      # Tensor parallel
PP=4      # Pipeline parallel
CP=1      # Context parallel

# LLaMA 70B configuration
MODEL_SIZE=70  # Billion parameters
HIDDEN_SIZE=8192
NUM_LAYERS=80
NUM_HEADS=64
SEQ_LENGTH=4096

# Training hyperparameters
MICRO_BATCH=1
GLOBAL_BATCH=1024
LR=3e-4

torchrun \
  --nproc_per_node=$GPUS_PER_NODE \
  --nnodes=$NNODES \
  pretrain_gpt.py \
  --tensor-model-parallel-size $TP \
  --pipeline-model-parallel-size $PP \
  --context-parallel-size $CP \
  --sequence-parallel \
  --num-layers $NUM_LAYERS \
  --hidden-size $HIDDEN_SIZE \
  --num-attention-heads $NUM_HEADS \
  --seq-length $SEQ_LENGTH \
  --max-position-embeddings $SEQ_LENGTH \
  --micro-batch-size $MICRO_BATCH \
  --global-batch-size $GLOBAL_BATCH \
  --lr $LR \
  --train-iters 100000 \
  --lr-decay-style cosine \
  --lr-warmup-iters 2000 \
  --weight-decay 0.1 \
  --clip-grad 1.0 \
  --bf16 \
  --use-mcore-models \
  --transformer-impl transformer_engine \
  --data-path /path/to/data \
  --vocab-file /path/to/vocab.json \
  --merge-file /path/to/merges.txt

Step 3: Launch distributed training

# Single node (8 GPUs)
bash train_llama_70b.sh

# Multi-node with SLURM
sbatch --nodes=8 --gpus-per-node=8 train_llama_70b.sh

Step 4: Monitor performance metrics

Key metrics to track:

Model FLOP Utilization (MFU): Target >40% on H100
Throughput: Tokens/sec/GPU
Memory usage: <80GB per GPU for 70B model
Loss: Should decrease steadily

Workflow 2: Configure Mixture of Experts (MoE) training

For sparse MoE models like Mixtral.

MoE Training:
- [ ] Step 1: Configure expert parallelism
- [ ] Step 2: Set MoE hyperparameters
- [ ] Step 3: Launch training with EP

Step 1: Configure expert parallelism

# Mixtral 8x7B example
TENSOR_PARALLEL=2
PIPELINE_PARALLEL=1
EXPERT_PARALLEL=4  # Split 8 experts across 4 GPUs
DATA_PARALLEL=4

TOTAL_GPUS=$((TENSOR_PARALLEL * PIPELINE_PARALLEL * EXPERT_PARALLEL * DATA_PARALLEL))
# = 2 * 1 * 4 * 4 = 32 GPUs

Step 2: Set MoE hyperparameters

torchrun \
  --nproc_per_node=8 \
  pretrain_gpt.py \
  --tensor-model-parallel-size 2 \
  --pipeline-model-parallel-size 1 \
  --expert-model-parallel-size 4 \
  --num-experts 8 \
  --moe-router-topk 2 \
  --moe-router-load-balancing-type aux_loss \
  --moe-aux-loss-coeff 0.01 \
  --hidden-size 4096 \
  --num-layers 32 \
  --num-attention-heads 32 \
  --seq-length 4096 \
  --max-position-embeddings 4096 \
  --bf16 \
  --use-mcore-models \
  --transformer-impl transformer_engine \
  --data-path /path/to/data \
  --vocab-file /path/to/vocab.json \
  --merge-file /path/to/merges.txt

Step 3: Launch training with EP

Expert parallelism distributes different experts across GPUs, reducing memory while maintaining capacity.

Memory without EP: 8 experts × 7B = 56GB per GPU
Memory with EP=4: 2 experts × 7B = 14GB per GPU
Savings: 75% memory reduction

Workflow 3: Optimize for maximum throughput

Achieve 47% MFU on H100.

Performance Optimization:
- [ ] Step 1: Enable Flash Attention
- [ ] Step 2: Use FP8 precision (H100)
- [ ] Step 3: Optimize micro-batch size
- [ ] Step 4: Tune parallelism degrees

Step 1: Enable optimizations

--use-mcore-models  # Use Megatron Core models
--transformer-impl transformer_engine  # Use Transformer Engine
--sequence-parallel  # Reduce activation memory (use with TP)

Step 2: Use FP8 precision (H100 only)

--fp8-hybrid  # FP8 mixed precision training
# Transformer Engine handles FP8 automatically

Result: 1.5-2x speedup on H100 vs BF16.

Step 3: Optimize micro-batch size

Find largest micro-batch that fits in memory:

# Start with 1, increase until OOM
for MBS in 1 2 4 8; do
  echo "Testing micro-batch-size=$MBS"
  torchrun ... --micro-batch-size $MBS
done

Typical values:

  • 7B model: 4-8
  • 70B model: 1-2
  • 405B model: 1

Step 4: Tune parallelism degrees

Rules of thumb:

Tensor Parallel: Use ≤8 (limited by NVLink within node)
Pipeline Parallel: Use for >70B models
Context Parallel: Use for sequences >8K tokens
Data Parallel: Fill remaining GPUs

Example 405B on 128 H100s:

TP=8 (1 node)
PP=8 (across nodes)
CP=2 (long sequences)
DP=1
Total = 8 × 8 × 2 × 1 = 128 GPUs

When to use vs alternatives

Use Megatron-Core when:

  • Training models >10B parameters
  • Need maximum efficiency (target >40% MFU)
  • Using NVIDIA GPUs (A100, H100)
  • Production training at scale
  • Want fine-grained parallelism control

Use alternatives instead:

  • PyTorch FSDP: Models <70B, simpler API, PyTorch native
  • DeepSpeed: Easier setup, good for <100B models
  • HuggingFace Accelerate: Prototyping, simpler workflows
  • LitGPT: Educational, single-file implementations

Common issues

Issue: Low GPU utilization (<30% MFU)

Causes:

  1. Micro-batch too small
  2. Too much parallelism overhead
  3. Not using Flash Attention

Fixes:

# Increase micro-batch
--micro-batch-size 4  # Was 1

# Enable optimizations
--use-flash-attn
--sequence-parallel

# Reduce TP if >8
--tensor-model-parallel-size 4  # Was 16

Issue: Out of memory

Reduce memory with:

--tensor-model-parallel-size 2  # Split model across GPUs
--recompute-granularity full  # Gradient checkpointing
--recompute-method block  # Checkpoint transformer blocks
--recompute-num-layers 1  # Checkpoint every layer

Or use CPU/NVMe offloading:

--cpu-optimizer  # Offload optimizer to CPU
--cpu-optimizer-type ADAM  # CPU Adam variant

Issue: Training slower than expected

Check:

  1. Network bottleneck: Ensure InfiniBand/NVLink enabled
  2. Pipeline bubbles: Use interleaved pipeline schedule
    --num-layers-per-virtual-pipeline-stage 2
    
  3. Data loading: Use fast data loader
    --dataloader-type cyclic
    

Issue: Diverging loss

Stabilize training:

--lr-warmup-iters 2000  # Longer warmup
--clip-grad 1.0  # Gradient clipping
--init-method-std 0.006  # Smaller init
--attention-dropout 0.0  # No dropout in attention
--hidden-dropout 0.0  # No dropout in FFN

Advanced topics

Parallelism strategies: See references/parallelism-guide.md for detailed comparison of TP/PP/DP/CP/EP with performance analysis and when to use each.

Performance benchmarks: See references/benchmarks.md for MFU numbers across different model sizes and GPU configurations.

Production configurations: See references/production-examples.md for real-world setups from LLaMA 3 405B, Nemotron-4 340B, and DeepSeek-V3 671B.

Training recipes: See references/training-recipes.md for complete hyperparameter configurations for GPT/LLaMA/Mixtral architectures.

Hardware requirements

  • GPU: NVIDIA Ampere+ (A100, H100, B200)
    • Turing works but slower
    • FP8 requires Hopper/Ada/Blackwell
  • Network: InfiniBand or 400Gb+ Ethernet for multi-node
  • Memory per GPU:
    • 7B model: 40GB+
    • 70B model: 80GB (with TP=4)
    • 405B model: 80GB (with TP=8, PP=8)
  • Storage: Fast NVMe for checkpoints (1TB+ for 70B+ models)

Resources

When not to use it

  • Small models that fit on single GPUs
  • Prototyping without the need for distributed infrastructure

Prerequisites

NVIDIA Megatron-CoretorchrunMultiple GPUs

Limitations

  • High infrastructure requirement for full parallelism usage
  • Steep configuration learning curve for distributed settings

How it compares

Handles the complex boilerplate of sharding models across thousands of GPUs, achieving higher FLOP efficiency than standard frameworks.

Compared to similar skills

training-llms-megatron side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
training-llms-megatron (this skill)17moReviewAdvanced
mlflow37moReviewIntermediate
modal-serverless-gpu17moReviewIntermediate
hugging-face-model-trainer16moReviewAdvanced

Try saying

Example prompts that trigger this skill in your AI assistant.

software-architecture

davila7

Guide for quality focused software architecture. This skill should be used when users want to write code, design architecture, analyze code, in any case that relates to software development.

333868

planning-with-files

davila7

Implements Manus-style file-based planning for complex tasks. Creates task_plan.md, findings.md, and progress.md. Use when starting complex multi-step tasks, research projects, or any task requiring >5 tool calls.

233106

telegram-bot-builder

davila7

Expert in building Telegram bots that solve real problems - from simple automation to complex AI-powered bots. Covers bot architecture, the Telegram Bot API, user experience, monetization strategies, and scaling bots to thousands of users. Use when: telegram bot, bot api, telegram automation, chat bot telegram, tg bot.

106130

scroll-experience

davila7

Expert in building immersive scroll-driven experiences - parallax storytelling, scroll animations, interactive narratives, and cinematic web experiences. Like NY Times interactives, Apple product pages, and award-winning web experiences. Makes websites feel like experiences, not just pages. Use when: scroll animation, parallax, scroll storytelling, interactive story, cinematic website.

101142

humanizer

davila7

Remove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, negative parallelisms, and excessive conjunctive phrases. Credits: Original skill by @blader - https://github.com/blader/humanizer

90175

game-development

davila7

Game development orchestrator. Routes to platform-specific skills based on project needs.

70195

You might also like

mlflow

davila7

Track ML experiments, manage model registry with versioning, deploy models to production, and reproduce experiments with MLflow - framework-agnostic ML lifecycle platform

322

modal-serverless-gpu

davila7

Serverless GPU cloud platform for running ML workloads. Use when you need on-demand GPU access without infrastructure management, deploying ML models as APIs, or running batch jobs with automatic scaling.

17

hugging-face-model-trainer

patchy631

This skill should be used when users want to train or fine-tune language models using TRL (Transformer Reinforcement Learning) on Hugging Face Jobs infrastructure. Covers SFT, DPO, GRPO and reward modeling training methods, plus GGUF conversion for local deployment. Includes guidance on the TRL Jobs package, UV scripts with PEP 723 format, dataset preparation and validation, hardware selection, cost estimation, Trackio monitoring, Hub authentication, and model persistence. Should be invoked for tasks involving cloud GPU training, GGUF conversion, or when users mention training on Hugging Face Jobs without local GPU setup.

13

tensorrt-llm

davila7

Optimizes LLM inference with NVIDIA TensorRT for maximum throughput and lowest latency. Use for production deployment on NVIDIA GPUs (A100/H100), when you need 10-100x faster inference than PyTorch, or for serving models with quantization (FP8/INT4), in-flight batching, and multi-GPU scaling.

10

flash

MachinaIO

runpod-flash SDK and CLI for deploying AI workloads on Runpod serverless GPUs/CPUs.

00

building-automl-pipelines

jeremylongshore

Build automated machine learning pipelines, including feature engineering, model selection, and performance evaluation.

688

Search skills

Search the agent skills registry