alignment-handbook/scripts/README.md

Scripts to Train and Evaluate Chat Models

Fine-tuning

In the handbook, we provide three main ways to align LLMs for chat:

• Full fine-tuning on a multi-GPU machine with DeepSpeed ZeRO-3 (tested on an 8 x A100 (80GB) node).
• LoRA fine-tuning on a single consumer 24GB GPU (tested on a RTX 4090).
• LoRA fine-tuning on a multi-GPU machine with DeepSpeed ZeRO-3 (tested on a 2 x A100s (80GB)).

In practice, we find comparable performance for both full and LoRA fine-tuning, with the latter having the advantage of producing small adapter weights that are fast to upload and download from the Hugging Face Hub. Here's the two general commands to fine-tune your models:

# Full training with ZeRO-3 on 8 GPUs
ACCELERATE_LOG_LEVEL=info accelerate launch --config_file recipes/accelerate_configs/deepspeed_zero3.yaml scripts/run_{task}.py recipes/{model_name}/{task}/config_full.yaml

# LoRA training on a single GPU
ACCELERATE_LOG_LEVEL=info accelerate launch --config_file recipes/accelerate_configs/multi_gpu.yaml --num_processes=1 scripts/run_{task}.py recipes/{model_name}/{task}/config_lora.yaml

# LoRA training with ZeRO-3 on two or more GPUs
ACCELERATE_LOG_LEVEL=info accelerate launch --config_file recipes/accelerate_configs/deepspeed_zero3.yaml --num_processes={num_gpus} scripts/run_{task}.py recipes/{model_name}/{task}/config_lora.yaml

Here {task} refers to type of training you wish to run (SFT, DPO, etc), while {model_name} refers to the choice of recipe in the recipes directory. For example, to replicate Zephyr-7B-β you can run:

# Step 1 - train SFT policy
ACCELERATE_LOG_LEVEL=info accelerate launch --config_file recipes/accelerate_configs/deepspeed_zero3.yaml scripts/run_sft.py recipes/zephyr-7b-beta/sft/config_full.yaml

# Step 2 - align with DPO
ACCELERATE_LOG_LEVEL=info accelerate launch --config_file recipes/accelerate_configs/deepspeed_zero3.yaml scripts/run_dpo.py recipes/zephyr-7b-beta/dpo/config_full.yaml

** 💡 Tip:** If you scale up/down the number of GPUs, we recommend also scaling up the per-device batch size or number of gradient accumulation steps to keep the global batch size constant (and thus replicate our results).

By default, these scripts will push each model to your Hugging Face Hub username, i.e. {username}/{model_name}-{task}. You can override the parameters in each YAML config by appending them to the command as follows:

# Change batch size, number of epochs etc
ACCELERATE_LOG_LEVEL=info accelerate launch --config_file recipes/accelerate_configs/deepspeed_zero3.yaml scripts/run_{task}.py recipes/{model_name}/{task}/config_full.yaml --per_device_train_batch_size=42 --num_train_epochs=5

By default all training metrics are logged with TensorBoard. If you have a Weights and Biases account and are logged in, you can view the training metrics by appending --report_to=wandb, e.g.

ACCELERATE_LOG_LEVEL=info accelerate launch --config_file recipes/accelerate_configs/deepspeed_zero3.yaml scripts/run_{task}.py recipes/{model_name}/{task}/config_full.yaml --report_to=wandb

Launching jobs on a Slurm cluster

If you have access to a Slurm cluster, we provide a recipes/launch.slurm script that will automatically queue training jobs for you. Here's how you can use it:

sbatch --job-name=handbook_{task} --nodes=1 recipes/launch.slurm {model_name} {task} {precision} {accelerator}

Here {model_name} and {task} are defined as above, while {precision} refers to the type of training (full vs LoRA) and {accelerator} refers to the choice of 🤗 Accelerate config in recipes/accelerate_configs. Here's a concrete example to run SFT on 1 node of 8 GPUs:

sbatch --job-name=handbook_sft --nodes=1 recipes/launch.slurm zephyr-7b-beta sft full deepspeed_zero3

You can scale the number of nodes by increasing the --nodes flag.

⚠️ Note: the configuration in recipes/launch.slurm is optimised for the Hugging Face Compute Cluster and may require tweaking to be adapted to your own compute nodes.

Fine-tuning on custom datasets

Under the hood, each training script uses the get_datasets() function which allows one to easily combing multiple datasets with varying proportions. For instance, this is how one can specify multiple datasets and which splits to combine in one of the YAML configs:

datasets_mixer:
    dataset_1: 0.5  # Use 50% of the training examples
    dataset_2: 0.66 # Use 66% of the training examples
    dataset_3: 0.10 # Use 10% of the training examples
dataset_splits:
- train_x           # Samples from each train split
- test_x            # Test splits aren't sampled

If you want to fine-tune on your own datasets, the main thing to keep in mind is how the chat templates are applied to the dataset blend. Since each task (SFT, DPO, etc), requires a different format, we assume the datasets have the following columns:

SFT

• messages: A list of dicts in the form {"role": "{role}", "content": {content}}.
• See ultrachat_200k for an example.

DPO

• chosen: A list of dicts in the form {"role": "{role}", "content": {content}} corresponding to the preferred dialogue.
• rejected: A list of dicts in the form {"role": "{role}", "content": {content}} corresponding to the dispreferred dialogue.
• See ultrafeedback_binarized for an example.

We also find it useful to include dedicated splits per task in our datasets, so e.g. we have:

• {train,test}_sft: Splits for SFT training.
• {train,test}_gen: Splits for generation ranking like rejection sampling or PPO.
• {train,test}_prefs: Splits for preference modelling, like reward modelling or DPO.

If you format your dataset in the same way, our training scripts should work out of the box!