| name | training-llms-megatron |
| description | 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. |
| version | 1.0.0 |
| author | Orchestra Research |
| license | MIT |
| tags | ["Megatron-Core","Large-Scale Training","NVIDIA","Tensor Parallelism","Pipeline Parallelism","Model Parallelism","H100","Distributed Training","Production"] |
| dependencies | ["megatron-core","torch","apex","transformer-engine"] |
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 run --gpus all -it --rm nvcr.io/nvidia/pytorch:25.04-py3
pip install megatron-core
Simple distributed training:
torchrun --nproc_per_node=2 examples/run_simple_mcore_train_loop.py
./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 Size | GPUs | Tensor Parallel | Pipeline Parallel | Data Parallel | Context Parallel |
|---|
| 7B | 8 | 1 | 1 | 8 | 1 |
| 13B | 8 | 2 | 1 | 4 | 1 |
| 70B | 64 | 4 | 4 | 4 | 1 |
| 405B | 128 | 8 | 8 | 2 | 2 |
Step 2: Configure training hyperparameters
#!/bin/bash
GPUS_PER_NODE=8
NNODES=8
TP=4
PP=4
CP=1
MODEL_SIZE=70
HIDDEN_SIZE=8192
NUM_LAYERS=80
NUM_HEADS=64
SEQ_LENGTH=4096
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 $LR \
--train-iters 100000 \
--lr-decay-style cosine \
--lr-warmup-iters 2000 \
--weight-decay 0.1 \
--clip-grad 1.0 \
--bf16 \
--use-mcore-models \
--transformer-impl transformer_engine \
--data-path /path/to/data \
--vocab-file /path/to/vocab.json \
--merge-file /path/to/merges.txt
Step 3: Launch distributed training
bash train_llama_70b.sh
sbatch --nodes=8 --gpus-per-node=8 train_llama_70b.sh
Step 4: Monitor performance metrics
Key metrics to track:
Model FLOP Utilization (MFU): Target >40% on H100
Throughput: Tokens/sec/GPU
Memory usage: <80GB per GPU for 70B model
Loss: Should decrease steadily
Workflow 2: Configure Mixture of Experts (MoE) training
For sparse MoE models like Mixtral.
MoE Training:
- [ ] Step 1: Configure expert parallelism
- [ ] Step 2: Set MoE hyperparameters
- [ ] Step 3: Launch training with EP
Step 1: Configure expert parallelism
TENSOR_PARALLEL=2
PIPELINE_PARALLEL=1
EXPERT_PARALLEL=4
DATA_PARALLEL=4
TOTAL_GPUS=$((TENSOR_PARALLEL * PIPELINE_PARALLEL * EXPERT_PARALLEL * DATA_PARALLEL))
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
--transformer-impl transformer_engine
--sequence-parallel
Step 2: Use FP8 precision (H100 only)
--fp8-hybrid
Result: 1.5-2x speedup on H100 vs BF16.
Step 3: Optimize micro-batch size
Find largest micro-batch that fits in memory:
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:
- Micro-batch too small
- Too much parallelism overhead
- Not using Flash Attention
Fixes:
--micro-batch-size 4
--use-flash-attn
--sequence-parallel
--tensor-model-parallel-size 4
Issue: Out of memory
Reduce memory with:
--tensor-model-parallel-size 2
--recompute-granularity full
--recompute-method block
--recompute-num-layers 1
Or use CPU/NVMe offloading:
--cpu-optimizer
--cpu-optimizer-type ADAM
Issue: Training slower than expected
Check:
- Network bottleneck: Ensure InfiniBand/NVLink enabled
- Pipeline bubbles: Use interleaved pipeline schedule
--num-layers-per-virtual-pipeline-stage 2
- Data loading: Use fast data loader
--dataloader-type cyclic
Issue: Diverging loss
Stabilize training:
--lr-warmup-iters 2000
--clip-grad 1.0
--init-method-std 0.006
--attention-dropout 0.0
--hidden-dropout 0.0
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