Important

You are viewing the NeMo 2.0 documentation. This release introduces significant changes to the API and a new library, NeMo Run. We are currently porting all features from NeMo 1.0 to 2.0. For documentation on previous versions or features not yet available in 2.0, please refer to the NeMo 24.07 documentation.

Llama#

Meta’s Llama builds on the general transformer decoder framework with some key additions such as pre-normalization, SwiGLU activations, and Rotary Positional Embeddings (RoPE). More information is available in the companion paper “Llama: Open and Efficient Foundation Language Models”. With a wide variety of model sizes - Llama has options for every inference budget.

We provide pre-defined recipes for pretraining and finetuning a Llama 3 model in two sizes: 8B and 70B, as well as Llama 3.1 model in three sizes: 8B, 70B and 405B. The recipes use NeMo 2.0 and NeMo-Run. These recipes configure a run.Partial for one of the nemo.collections.llm api functions introduced in NeMo 2.0. The recipes are hosted in the following files: llama3_8b, llama3_70b, llama31_8b, llama31_70b, llama31_405b.

NeMo 2.0 Pretraining Recipes#

Note

The pretraining recipes use the MockDataModule for the data argument. You are expected to replace the MockDataModule with your custom dataset.

We provide an example below on how to invoke the default recipe and override the data argument:

from nemo.collections import llm

pretrain = llm.llama3_8b.pretrain_recipe(
    name="llama3_8b_pretraining",
    ckpt_dir=f"/path/to/checkpoints",
    num_nodes=1,
    num_gpus_per_node=8,
)

# # To override the data argument
# dataloader = a_function_that_configures_your_custom_dataset(
#     gbs=gbs,
#     mbs=mbs,
#     seq_length=pretrain.model.config.seq_length,
# )
# pretrain.data = dataloader

NeMo 2.0 Finetuning Recipes#

Note

The finetuning recipes use the SquadDataModule for the data argument. You can replace the SquadDataModule with your custom dataset.

To import the HF model and convert to NeMo 2.0 format, run the following command (this only needs to be done once)

from nemo.collections import llm
llm.import_ckpt(model=llm.LlamaModel(llm.Llama3Config8B()), source='hf://meta-llama/Meta-Llama-3-8B')

By default, the non-instruct version of the model is loaded. To load a different model, set finetune.resume.restore_config.path=nemo://<hf model id> or finetune.resume.restore_config.path=<local model path>

We provide an example below on how to invoke the default recipe and override the data argument:

from nemo.collections import llm

recipe = llm.llama3_8b.finetune_recipe(
    name="llama3_8b_finetuning",
    ckpt_dir=f"/path/to/checkpoints",
    num_nodes=1,
    num_gpus_per_node=8,
    peft_scheme='lora',  # 'lora', 'none'
    packed_sequence=False,
)

# # To override the data argument
# dataloader = a_function_that_configures_your_custom_dataset(
#     gbs=gbs,
#     mbs=mbs,
#     seq_length=recipe.model.config.seq_length,
# )
# recipe.data = dataloader

By default, the finetuning recipe will run LoRA finetuning with LoRA applied to all linear layers in the language model. To finetune the entire model without LoRA, set peft_scheme='none' in the recipe argument.

To finetune with sequence packing for a higher throughput, set packed_sequence=True. Note that you may need to tune the global batch size in order to achieve similar convergence.

Note

The configuration in the recipes is done using the NeMo-Run run.Config and run.Partial configuration objects. Please review the NeMo-Run documentation to learn more about its configuration and execution system.

Once you have your final configuration ready, you can execute it on any of the NeMo-Run supported executors. The simplest is the local executor, which just runs the pretraining locally in a separate process. You can use it as follows:

import nemo_run as run

run.run(pretrain, executor=run.LocalExecutor())

Additionally, you can also run it directly in the same Python process as follows:

run.run(pretrain, direct=True)

A comprehensive list of pretraining recipes that we currently support or plan to support soon is provided below for reference:

Recipe

Status

Llama 3 8B

Yes

Llama 3 8B FP8

N/A

Llama 3 8B 16k seq length

Yes

Llama 3 8B 64k seq length

Yes

Llama 3 70B

Yes

Llama 3 70B FP8

N/A

Llama 3 70B 16k seq length

Yes

Llama 3 70B 64k seq length

Yes

Llama 3.1 8B

Yes

Llama 3.1 70B

Yes

Llama 3.1 405B

Yes
On this page