NeMo 2.0 Finetuning Recipes#

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

fromnemo.collectionsimportllmif__name__=="__main__":llm.import_ckpt(model=llm.Gemma2Model(llm.Gemma2Config2B()),source='hf://google/gemma-2-2b')

By default, the non-instruct version of the model is loaded. To load a different model, setfinetune.resume.restore_config.path=nemo://<hfmodelid> orfinetune.resume.restore_config.path=<localmodelpath>

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

fromnemo.collectionsimportllmrecipe=llm.gemma2_2b.finetune_recipe(name="gemma2_2b_finetuning",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, setpeft_scheme='none' in the recipe argument.

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

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:

importnemo_runasrunrun.run(recipe,executor=run.LocalExecutor())

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

run.run(recipe,direct=True)