NVFP4 Pretraining with Megatron Bridge

Step 1
Set up the environment

The recommended way to run Megatron-Bridge on DGX Station is through the NeMo Framework container, which includes Megatron-Bridge, Megatron-Core, Transformer Engine, and all CUDA dependencies pre-installed. Running outside the container is not supported in this playbook — the NVFP4 kernels rely on the exact Transformer Engine / CUDA versions shipped inside the image.

The training recipe fetches the Llama 3 8B architecture config from HuggingFace, so export a HuggingFace token with access to meta-llama/Meta-Llama-3-8B before launching the container:

Bash

export HF_TOKEN=<YOUR_HF_TOKEN>

Bash

git clone https://github.com/NVIDIA/dgx-spark-playbooks
cd dgx-spark-playbooks/nvidia/station-nvfp4-pretraining/assets

# Use the latest nemo tag
export TAG=26.04

# This training primarily works on GB300 GPU
GB300_DEVICE=$(nvidia-smi --query-gpu=index,name --format=csv,noheader | awk -F', ' '/GB300/ {print $1; exit}')

docker run --rm -it \
  --gpus device=${GB300_DEVICE} \
  --ipc host \
  --ulimit memlock=-1 \
  --ulimit stack=67108864 \
  -e HF_TOKEN="$HF_TOKEN" \
  -v "$HOME/.cache/huggingface:/root/.cache/huggingface" \
  -v "$(pwd):/workdir" \
  -w /workdir \
  --entrypoint bash \
  nvcr.io/nvidia/nemo:${TAG}

All subsequent torchrun / python commands in this playbook are meant to be executed from the shell inside this container.

Step 2
Review the pretraining script

The pretraining script can be found at pretrain_llama.py. The key piece is the NVFP4 precision config, built on top of Megatron-Bridge's prebuilt bf16_with_nvfp4_mixed recipe:

Python

from megatron.bridge.training.mixed_precision import bf16_with_nvfp4_mixed

def nvfp4_mixed_precision():
    cfg = bf16_with_nvfp4_mixed()
    cfg.first_last_layers_bf16 = True
    cfg.num_layers_at_start_in_bf16 = 0
    cfg.num_layers_at_end_in_bf16 = 4
    return cfg

bf16_with_nvfp4_mixed() already sets fp8="e4m3" and fp8_recipe="nvfp4" under the hood; we just toggle the layer-pinning knobs on top:

Last 4 layers in BF16 (num_layers_at_end_in_bf16=4) for training stability (adjustable per model)
No start-layer pinning (num_layers_at_start_in_bf16=0) — last-layer stability is usually enough

NOTE

The script uses llama3_8b_pretrain_config() which defaults to context_parallel_size=2. The script overrides this to context_parallel_size=1 for single-GPU runs. If you swap in a larger recipe (e.g. nemotron_3_nano_pretrain_config, which defaults to TP=4), you must either launch torchrun --nproc_per_node=4 on a 4-GPU node or override config.model.tensor_model_parallel_size = 1 before calling pretrain(...), or you will hit: AssertionError: world size (1) is not divisible by total_model_size (...tensor_model_parallel_size=4 * ...).

Step 3
Launch NVFP4 pre-training

Launch a short training run with mock data and tee the output to a log file so you can inspect VRAM and per-iteration latency afterwards:

Bash

torchrun --nproc_per_node=1 pretrain_llama.py > nvfp4.log 2>&1

Expected output (see nvfp4.log):

Model initialization logs and a Theoretical memory footprints: weight and optimizer=... line
Iteration progress printed every step (log_interval=1), e.g. iteration 10/50 | ... elapsed time per iteration (ms): ... | lm loss: ...
A [Rank 0] ... memory (GB) | mem-max-reserved-gigabytes: ... line — this is your peak VRAM
A checkpoint saved to /workdir/nemo_experiments/default/checkpoints

If the run finishes with EXIT=0 (or no traceback), your NVFP4 pretraining setup is working.

Step 4
Compare with BF16 baseline

Run the same script with --disable-fp4 to establish a BF16 baseline, again logging to a file:

Bash

# Remove the prior checkpoint directory so the two runs don't interfere
rm -rf nemo_experiments

torchrun --nproc_per_node=1 pretrain_llama.py --disable-fp4 > bf16.log 2>&1

To compare the two runs on latency and throughput, grep the per-iteration lines out of each log:

Bash

grep -E "elapsed time per iteration|MODEL_TFLOP" nvfp4.log
grep -E "elapsed time per iteration|MODEL_TFLOP" bf16.log

Each step prints two lines:

Step Time : 5.39s GPU utilization: 2347.0MODEL_TFLOP/s/GPU — step latency and throughput
iteration 10/50 | ... elapsed time per iteration (ms): 5390 | ... lm loss: ... — same latency in ms plus loss

Iteration 10 includes one-time CUDA-graph/compile overhead, so average iterations 20–50 for a fair per-step latency number.

Measuring peak VRAM (from `nvidia-smi`)

Megatron's in-log memory numbers (mem-max-reserved-gigabytes) reflect PyTorch's caching-allocator reservation, which can drift from what the device actually holds. For an accurate read, watch nvidia-smi live from a second shell while training runs:

Bash

watch -n 1 nvidia-smi

See the measured numbers in overview.md for expected VRAM and latency on 1× GB300 with Llama 3.1 8B.

Step 5
Script arguments

pretrain_llama.py accepts the following arguments:

Argument	Type	Default	Description
`--disable-fp4`	flag	off	Disable NVFP4; use plain BF16 mixed precision as a baseline
`--train-iters`	int	50	Number of training iterations
`--warmup-iters`	int	2	Number of warmup iterations
`--global-batch-size`	int	64	Global batch size
`--micro-batch-size`	int	4	Micro batch size (drives peak VRAM; increase to use more memory)
`--seq-length`	int	4096	Sequence length

Example combining several flags:

Bash

torchrun --nproc_per_node=1 pretrain_llama.py \
    --train-iters 50 --warmup-iters 2 \
    --global-batch-size 64 --micro-batch-size 4 --seq-length 4096

Step 6
Point to real data

To train on your own dataset, modify the config in the script:

Python

config = llama3_8b_pretrain_config()
config.dataset.data_path = ["/path/to/your/preprocessed/dataset"]
config.train.train_iters = 5000
config.train.global_batch_size = 256
config.train.micro_batch_size = 2

Megatron-Bridge expects preprocessed data in Megatron format. See the Megatron-Bridge data preparation guide for details.

Step 7
Cleanup

Remove checkpoints and log files generated by the runs:

Bash

rm -rf nemo_experiments/ nvfp4.log bf16.log

Then exit the container shell (exit) — the --rm flag in Step 1 deletes it automatically.

References

Step 1
Set up the environment

The training recipe fetches the Llama 3 8B architecture config from HuggingFace, so export a HuggingFace token with access to meta-llama/Meta-Llama-3-8B before launching the container:

Bash

export HF_TOKEN=<YOUR_HF_TOKEN>

Bash

git clone https://github.com/NVIDIA/dgx-spark-playbooks
cd dgx-spark-playbooks/nvidia/station-nvfp4-pretraining/assets

# Use the latest nemo tag
export TAG=26.04

# This training primarily works on GB300 GPU
GB300_DEVICE=$(nvidia-smi --query-gpu=index,name --format=csv,noheader | awk -F', ' '/GB300/ {print $1; exit}')

docker run --rm -it \
  --gpus device=${GB300_DEVICE} \
  --ipc host \
  --ulimit memlock=-1 \
  --ulimit stack=67108864 \
  -e HF_TOKEN="$HF_TOKEN" \
  -v "$HOME/.cache/huggingface:/root/.cache/huggingface" \
  -v "$(pwd):/workdir" \
  -w /workdir \
  --entrypoint bash \
  nvcr.io/nvidia/nemo:${TAG}

All subsequent torchrun / python commands in this playbook are meant to be executed from the shell inside this container.

Step 2
Review the pretraining script

The pretraining script can be found at pretrain_llama.py. The key piece is the NVFP4 precision config, built on top of Megatron-Bridge's prebuilt bf16_with_nvfp4_mixed recipe:

Python

from megatron.bridge.training.mixed_precision import bf16_with_nvfp4_mixed

def nvfp4_mixed_precision():
    cfg = bf16_with_nvfp4_mixed()
    cfg.first_last_layers_bf16 = True
    cfg.num_layers_at_start_in_bf16 = 0
    cfg.num_layers_at_end_in_bf16 = 4
    return cfg

bf16_with_nvfp4_mixed() already sets fp8="e4m3" and fp8_recipe="nvfp4" under the hood; we just toggle the layer-pinning knobs on top:

Last 4 layers in BF16 (num_layers_at_end_in_bf16=4) for training stability (adjustable per model)
No start-layer pinning (num_layers_at_start_in_bf16=0) — last-layer stability is usually enough

NOTE

Step 3
Launch NVFP4 pre-training

Launch a short training run with mock data and tee the output to a log file so you can inspect VRAM and per-iteration latency afterwards:

Bash

torchrun --nproc_per_node=1 pretrain_llama.py > nvfp4.log 2>&1

Expected output (see nvfp4.log):

Model initialization logs and a Theoretical memory footprints: weight and optimizer=... line
Iteration progress printed every step (log_interval=1), e.g. iteration 10/50 | ... elapsed time per iteration (ms): ... | lm loss: ...
A [Rank 0] ... memory (GB) | mem-max-reserved-gigabytes: ... line — this is your peak VRAM
A checkpoint saved to /workdir/nemo_experiments/default/checkpoints

If the run finishes with EXIT=0 (or no traceback), your NVFP4 pretraining setup is working.

Step 4
Compare with BF16 baseline

Run the same script with --disable-fp4 to establish a BF16 baseline, again logging to a file:

Bash

# Remove the prior checkpoint directory so the two runs don't interfere
rm -rf nemo_experiments

torchrun --nproc_per_node=1 pretrain_llama.py --disable-fp4 > bf16.log 2>&1

To compare the two runs on latency and throughput, grep the per-iteration lines out of each log:

Bash

grep -E "elapsed time per iteration|MODEL_TFLOP" nvfp4.log
grep -E "elapsed time per iteration|MODEL_TFLOP" bf16.log

Each step prints two lines:

Step Time : 5.39s GPU utilization: 2347.0MODEL_TFLOP/s/GPU — step latency and throughput
iteration 10/50 | ... elapsed time per iteration (ms): 5390 | ... lm loss: ... — same latency in ms plus loss

Iteration 10 includes one-time CUDA-graph/compile overhead, so average iterations 20–50 for a fair per-step latency number.

Measuring peak VRAM (from `nvidia-smi`)

Bash

watch -n 1 nvidia-smi

See the measured numbers in overview.md for expected VRAM and latency on 1× GB300 with Llama 3.1 8B.

Step 5
Script arguments

pretrain_llama.py accepts the following arguments:

Argument	Type	Default	Description
`--disable-fp4`	flag	off	Disable NVFP4; use plain BF16 mixed precision as a baseline
`--train-iters`	int	50	Number of training iterations
`--warmup-iters`	int	2	Number of warmup iterations
`--global-batch-size`	int	64	Global batch size
`--micro-batch-size`	int	4	Micro batch size (drives peak VRAM; increase to use more memory)
`--seq-length`	int	4096	Sequence length

Example combining several flags:

Bash

torchrun --nproc_per_node=1 pretrain_llama.py \
    --train-iters 50 --warmup-iters 2 \
    --global-batch-size 64 --micro-batch-size 4 --seq-length 4096

Step 6
Point to real data

To train on your own dataset, modify the config in the script:

Python

config = llama3_8b_pretrain_config()
config.dataset.data_path = ["/path/to/your/preprocessed/dataset"]
config.train.train_iters = 5000
config.train.global_batch_size = 256
config.train.micro_batch_size = 2

Megatron-Bridge expects preprocessed data in Megatron format. See the Megatron-Bridge data preparation guide for details.

Step 7
Cleanup

Remove checkpoints and log files generated by the runs:

Bash

rm -rf nemo_experiments/ nvfp4.log bf16.log

Then exit the container shell (exit) — the --rm flag in Step 1 deletes it automatically.

NVFP4 Pretraining with Megatron Bridge

Step 1
Set up the environment

Step 2
Review the pretraining script

Step 3
Launch NVFP4 pre-training

Step 4
Compare with BF16 baseline

Measuring peak VRAM (from `nvidia-smi`)

Step 5
Script arguments

Step 6
Point to real data

Step 7
Cleanup

References

Resources

NVFP4 Pretraining with Megatron Bridge

Step 1
Set up the environment

Step 2
Review the pretraining script

Step 3
Launch NVFP4 pre-training

Step 4
Compare with BF16 baseline

Measuring peak VRAM (from `nvidia-smi`)

Step 5
Script arguments

Step 6
Point to real data

Step 7
Cleanup

References

Resources

NVFP4 Pretraining with Megatron Bridge

Step 1Set up the environment

Step 2Review the pretraining script

Step 3Launch NVFP4 pre-training

Step 4Compare with BF16 baseline

Measuring peak VRAM (from nvidia-smi)

Step 5Script arguments

Step 6Point to real data

Step 7Cleanup

References

Resources

NVFP4 Pretraining with Megatron Bridge

Step 1Set up the environment

Step 2Review the pretraining script

Step 3Launch NVFP4 pre-training

Step 4Compare with BF16 baseline

Measuring peak VRAM (from nvidia-smi)

Step 5Script arguments

Step 6Point to real data

Step 7Cleanup

References

Resources

Step 1
Set up the environment

Step 2
Review the pretraining script

Step 3
Launch NVFP4 pre-training

Step 4
Compare with BF16 baseline

Measuring peak VRAM (from `nvidia-smi`)

Step 5
Script arguments

Step 6
Point to real data

Step 7
Cleanup

Step 1
Set up the environment

Step 2
Review the pretraining script

Step 3
Launch NVFP4 pre-training

Step 4
Compare with BF16 baseline

Measuring peak VRAM (from `nvidia-smi`)

Step 5
Script arguments

Step 6
Point to real data

Step 7
Cleanup