Parameter Efficient Fine-Tuning (PEFT)
Run PEFT with NeMo Launcher
To run PEFT update conf/config.yaml
:
defaults:
- peft: Mistral/squad
stages:
- peft
Execute launcher pipeline: python3 main.py
Configuration
Default configurations for PEFT with squad can be found in conf/peft/mistral/squad.yaml
.
Fine-tuning configuration is divided into four sections run
, trainer
, exp_manger
and model
.
run:
name: peft_mistral_7b
time_limit: "04:00:00"
dependency: "singleton"
convert_name: convert_nemo
model_train_name: Mistral_7b
convert_dir: ${base_results_dir}/${peft.run.model_train_name}/${peft.run.convert_name}
task_name: "squad"
results_dir: ${base_results_dir}/${.model_train_name}/peft_${.task_name}
Set the number of nodes and devices for fine-tuning:
trainer:
num_nodes: 1
devices: 8
model:
restore_from_path: ${peft.run.convert_dir}/results/megatron_mistral.nemo
restore_from_path
sets the path to the .nemo
checkpoint to run fine-tuning.
peft_scheme
sets the fine-tuning scheme to be used. Supported schemes include: lora, adapter, ia3, ptuning.
Gated Model assets
Mistral’s tokenizer is hosted on Huggingface.com which requires login. In order to access the tokenizer assets, users are advised to prepend the HF_TOKEN=<token> environment variable to the nemo launcher invocation command.
In NeMo Laucher this can be achieved by appending “++env_vars.HF_TOKEN=<user-token” to the argument list.