BAdam

The implementation for BAdam: A Memory Efficient Full Parameter Training Method for Large Language Models. This paper presents an algorithm named BAdam, which finetunes Llama 2-7b and Llama 3-8B using a single RTX3090 with Adam's update rule and mixed precision training. The core idea of BAdam is to sequentially solve block coordinate optimization sub-problems. From the implementation perspective, the algorithm runs Adam's update on a small portition (usually one single transformer layer) of the parameters, thereby requires much less memory in comparison to full parameter Adam finetuning. Using BAdam only requires one line modification of the original code.

Method	Minimum Memory	Actual Memory Cost (Llama 3-8B)	Actual Memory Cost (Llama 2-7B)
Adam	$18M$	144 GB+	122.8 GB+
BAdam	$2M + \frac{16M}{D}$	23.5 GB	21.8 GB

Table 1: Comparison of Methods. $M$ stands for the number of model's parameters in billion and $D$ is the number of blocks used in BAdam. See Table 4 in paper for detailed analysis on memory consumption.

Method	Llama 3-8b	Llama 2-7b
Pretrained model	5.46	3.93
LoRA	6.41	5.05
BAdam	6.67	5.21

Table 2: MT bench score. The model is instruction finetuned on Alpaca-GPT4 dataset using a single RTX3090. BAdam consistently outperforms LoRA in MT bench under various evaluation models.

One can also apply BAdam for larger models with size such as 13B, 22B, 30B, and 70B. The memory consumption can be estimated to be $2M + \frac{16M}{D}$ (GB), plus some additional memory consumption for gradient checkpointed activations and system use like PyTorch's pre-allocation, etc (minor part).

Change log

[24/04/16] Our algorithm has been added to LLaMA-Factory. We would like to express our gratitude to their efforts on integrating BAdam!

[24/04/12] Add LoRA module detection. Make BlockOptimizer compatible with lr scheduler.

Environment Setup
Usage of BAdam
Run Paper Experiment
- Llama 3-8B and Llama 2-7B on Alpaca-GPT4
- RoBERTa-large on superGLUE

Setup

To install BAdam from Pypi, one can run:

pip install badam

One may also choose to build from source by the following steps:

git clone git@github.com:Ledzy/BAdam.git
cd BAdam
pip install -e .

For those who are interested in reproducing the results in paper, please follow the steps below to setup environment:

conda create -n badam python=3.10
conda activate badam
pip install -r requirements.txt

Usage of BAdam

Partition by Transfomer Layers

BAdam uses mixed-precision training, make sure that the model is loaded in float16 precision for memory saving. To use BAdam, one can simply add one line of code that wraps the original optimizer.

from badam import BlockOptimizer

# before training, add this line to wrap the original optimizer
optimizer = BlockOptimizer(
    base_optimizer=original_optimizer, # can be any torch.Optimizer
    named_parameters_list=list(model.named_parameters()), 
    switch_block_every=100, # switch to the new block every 50 updates, the $K$ Adam steps in paper. It can be set adaptively by $K = n/(BD)$, where $n$ is the number of training data points, $B$ is the batch size, and $D$ is the number of blocks in BAdam; see "Hyperparameter Suggestion" section for a detailed explaination about setting this hyperparameter. 
    switch_mode="random", # update order of blocks, one can choose "random" (random reshuffling update order), "ascending" (update from input layer to output layer), or "descending" (update from output layer to input layer). The default is "random".
    verbose=2 # information level, will print trainable parameters when setting to 2
)

The above code automatically creates the block list according to model.named_parameters. Specifically, it treates each transformer layer as a single block. For instance, for the Llama 2-7B, the block partition ($D = 32$) will be

block 1: model.layers.0.
block 2: model.layers.1.
...
block 32: model.layers.31.

By default, the embedding layer and language modeling head is not included in the training blocks. One can add them as two additional blocks by setting include_embedding=True, include_lm_head=True. One can also specify their own block list for the block optimizer. This can be achieved by adjusting theblock_prefix_list argument. For instance, the following code snippet divide each layer into blocks, which helps further reduce the memory cost:

block_prefix_list = []
for i in range(31):
    layer_prefix = [
        [f"model.layers.{i}.self_attn.q_proj."],
        [f"model.layers.{i}.self_attn.k_proj."],
        [f"model.layers.{i}.self_attn.v_proj."],
        [f"model.layers.{i}.self_attn.o_proj."],
        [f"model.layers.{i}.mlp.gate_proj."],
        [f"model.layers.{i}.mlp.up_proj."],
        [f"model.layers.{i}.mlp.down_proj."],
    ]
    block_prefix_list.extend(layer_prefix)

optimizer = BlockOptimizer(
    base_optimizer=original_optimizer,
    named_parameters_list=list(model.named_parameters_list), 
    switch_block_every=100,
    switch_mode="random",
    verbose=2,
    block_prefix_list=block_prefix_list # set the block list
)

Important Notes:

When setting block partition, one should be careful with the downstream task. Some tasks has randomly initialized classification layers, such as the SuperGLUE where the task_dict and pooler layers are randomly initialized. In this case, make sure to train these layers first, or set it to be trainable through the whole time. To set modules to be trainable through the whole training process, one can use active_modules argument, e.g., set active_modules=["model.task_dict.", "model.pooler."] when create the BlockOptimizer. Note that randomly initialized layers are usually the last layer, so updating these layers will only introduce negligible additional BP time. We thus suggest to always set the last classification layer to be trainable when the memory is permitted, if it is randomly initialized.
The parameters that are not included in block_prefix_list will be inactive (freezed) through the whole training procedure.
When setting prefix, it is suggested to include a . at the end. For example, it is preferred to use model.layers.1. instead of model.layers.1, as the later one includes the layer 10, 11, ..., 19 as well (since they have the same prefix).
Currently, all the experiments are conducted using a single 3090 GPU. Using this code in distributed training may exhibit unpredictable behaviors. For instance, when using pytorch DDP, the reducer for gradient synchronization are created when initializing the DDP optimizer. When switching to block where the reducer are not created, the block will NOT be updated as expected. The code version for distributed training is currently under active development.

Partition by Parameter Ratio

Instead of partitioning block by the model's parameter, an alternative choice is to train all the parameters simultaneously with a fixed ratio. For instance, we can train 5% parameters of every transformer layer. Namely, each active block contains 5% parameters from every transformer layer. In this sense, the feature extractor of every layer are jointly trained, which may be preferred in certain scenarios. However, training a block consisting of parameters coming from all the transformer layers may lose partly the benefit of BP time saving of BAdam.

To do this, one can use the BlockOptimizerRatio:

from badam import BlockOptimizerRatio

optimizer = BlockOptimizerRatio(
    param_groups=param_groups, # param_group of torch.Optimizer, the same as the original optimizer
    named_parameters_list=list(self.model.named_parameters()),
    switch_every=100, # switch to the new block every 100 updates
    update_ratio=0.1, # ratio of trainable weight for each parameter
    mask_mode = "adjacent", # choices: ["adjacent", "scatter"], see Note below for more explanation
    lr=1e-6,
    betas=(0.9, 0.999), # betas for Adam update
    eps=1e-8, # eps of Adam update
)

Currently, the BlockOptimizerRatio only supports the Adam update. The repository is still under active development.

Notes:

The mask_mode indicates how should the trainable parameter distribute across a parameter. mask_mode=adjacent indicates that the trainable parameters are adjacent to each other, while mask_mode=scatter indicates that trainable parameters are randomly choosed from the weight. For instance, considering optimizing a $10 \times 10$ matrix with update_ratio=0.1, setting mask_mode=adjacent will let parameters of the same row be the same block, and mask_mode=scatter means randomly choose 10 trainable parameters from the matrix.
By default, BlockOptimizerRatio does not update embedding layer, since in principle the embedding vectors of the tokens that are included in the training samples should be updated, while randomly freeze embedding parameters makes the update imbalanced. One can set include_embedding=True to include it for experimental purpose.
For BlockOptimizerRatio, we notice that setting mask_mode = "adjacent" usually performs the best; we leave the study of mask_mode as a future work. The convergence speed is highly positively related to the update_ratio, so we suggest to choose it as high as possible when the memory is permitted.
The gradient and optimizer states are stored in sparse tensor format. The update rule is exactly the same as the BlockOptimizer: run Adam update on current active block for switch_every steps, and then switch to next block.
Currently, the operation of sparsifing the gradient causes noticable overhead, which inevitably slow down the training. We leave the acceleration as a future work.

Hyperparameter Suggestion

Choice of the switch_block_every. Compared to Adam, our BAdam only introduces one additional hyperparameter, i.e., the switch_block_every (the K Adam steps in paper). It determines how many Adam steps we perform for each active block before switching to the next one. Fortunately, this hyperparameter can be set adaptively. Ideally, we expect to balance the data usage for each block in every epoch. This gives a natural choice of switch_block_every = $\frac{n}{BD}$ (rounding to the nearest integer if it is a fractional), where $n$ is the number of training data points, $B$ is the effective batch size, and $D$ is the number of blocks in BAdam. Using such a setting ensures that after one block-epoch, all the training data points are equally distributed to the $D$ blocks for training. Meanwhile, to achieve sufficient decrease for each block coordinate optimization subproblem and fully utilize the advantage of mixed precision training for reducing rounding error, the switch frequency should not be too small. We notice that set switch_block_every = $\max(\frac{n}{BD}, 50)$ usually yields fast convergence speed on both training loss and validation loss.

Run Paper Experiment

Llama 3-8B and Llama 2-7B on Alpaca-GPT4

Our implementation of finetuning Llama 3 and Llama 2 is based on Llama Factory. For the experiment of finetuning Llama-2 7b on Alpaca-GPT4 dataset, first change the working directory to llama:

cd llama-alpaca

Here is a sample command for running the code:

CUDA_VISIBLE_DEVICES=0 python src/train_bash.py \
    --stage sft \
    --model_name_or_path meta-llama/Llama-2-7b \
    --do_train \
    --dataset alpaca_gpt4_en \
    --template default \
    --finetuning_type block \
    --output_dir ./outputs/tmp \
    --overwrite_cache \
    --per_device_train_batch_size 2 \
    --per_device_eval_batch_size 2 \
    --gradient_accumulation_steps 8 \
    --lr_scheduler_type cosine \
    --logging_steps 1 \
    --save_steps 1000 \
    --val_size 500 \
    --eval_steps 20 \
    --evaluation_strategy steps \
    --learning_rate 1e-6 \
    --num_train_epochs 3 \
    --overwrite_output_dir \
    --plot_loss \
    --switch_block_every 100 \
    --switch_mode random \
    --bf16

To finetune Llama 3-8B, one can set --model_name_or_path meta-llama/Meta-Llama-3-8B. We use learning rate 1e-6 for both Llama 3-8B and Llama 2-7B.

One can also use Llama Factory to implement tuning Llama, as our BAdam is added to this factory.

Notes on arguments:

--stage: Currently we only implement the sft.
--finetuning_type: Options: (block, full, lora, sparse)
--switch_mode: How to order the block update. Options: (random, ascending, descending).
--switch_block_every: Switch block frequency; see "Hyperparameter Suggestion" for how to set this hyperparamter.
The above sample command is different from the hyperparameters settings in paper, while this version is more efficient. We will update our paper later.

RoBERTa-large on SuperGLUE

Our implementation for finetuning RoBERTa-large on superGLUE is based on jiant. To run the code, go to directory roberta-superglue first:

cd roberta-superglue

Before training the model, download the dataset using the following bash script. Adjust the script to download the required dataset.

EXP_DIR=./content/exp

python jiant/scripts/download_data/runscript.py \
    download \
    --tasks copa \
    --output_path ${EXP_DIR}/tasks

The finetuning command has the following form:

CUDA_VISIBLE_DEVICES=0 python badam_ft.py \
    --task_name boolq \
    --num_train_epochs 32 \
    --eval_every_steps 100 \
    --use_block_optim \
    --switch_every 100 \
    --switch_mode ascending \
    --train_batch_size 16 \
    --train_last_layer \
    --hf_pretrained_model_name FacebookAI/roberta-large