A PyTorch implementation of the model architecture of Pathways Language Model (PaLM): Scaling to 540 Billion Parameters for Breakthrough Performance. We take advantage of Colosssal-AI to exploit multiple optimization strategies, e.g. data parallelism, tensor parallelism, mixed precision & ZeRO, to scale the training to multiple GPUs.
You are very welcome to contribute in any way to help us enhance the usability of this project.
- Install requirements, e.g. Colosssal-AI, which is a Pytorch-based large-scale model training system with various efficient parallelization techniques.
pip install -r requirements.txt
- Use HuggingFace datasets to download Wikitext-2 dataset. The placeholder
/PATH/TO/DATA
is optional and is./wiki_dataset
by default.
python ./tools/download_wiki.py -o </PATH/TO/DATA>
- Download tokenizer files by calling the following command. The place holder
/PATH/TO/TOKENIZER/
is optional and is./token
by default.
bash ./tools/download_token.sh </PATH/TO/TOKENIZER/>
- Configure your settings in
CONFIG_FILE.py
like below. We also provide some examples in ./configs
SEQ_LENGTH = 512
BATCH_SIZE = 8
NUM_EPOCHS = 10
parallel = dict(
tensor=dict(mode='1d', size=2),
)
model = dict(type="palm_small")
- Set dataset & tokenizer paths
export DATA=</PATH/TO/DATA/>
export TOKENIZER=</PATH/TO/TOKENIZER/>
- Run
env OMP_NUM_THREADS=12 torchrun --nproc_per_node NUM_GPUS \
train.py --from_torch --config CONFIG_FILE.py
-
Run With Docker
Dockerfile is provided in this repository and you can run PaLM in Docker with the following commands.
# build docker image
docker build -t palm .
# exec training
docker run -ti --gpus all --rm palm \
torchrun --nproc_per_node NUM_GPUS \
train.py --from_torch --config CONFIG_FILE.py
The project has referred PaLM-Pytorch from lucidrains.