Spaces:
Running
Running
# How to use XTuner in HuggingFace training pipeline | |
## Quick run | |
1. step in `examples` | |
```shell | |
cd ./examples | |
``` | |
2. run training scripts | |
```shell | |
# qlora-training internlm-7b with alpaca dataset | |
python train_qlora_hf.py --model_name_or_path internlm/internlm-7b --dataset_name_or_path tatsu-lab/alpaca | |
``` | |
`--model_name_or_path`: specify the model name or path to train. | |
`--dataset_name_or_path`: specify the dataset name or path to use. | |
## How to customize your experiment | |
XTuner APIs are compatible with the usage of HuggingFace's transformers. | |
If you want to customize your experiment, you just need to pass in your hyperparameters like HuggingFace. | |
``` | |
# training example | |
python train_qlora_hf.py \ | |
# custom training args | |
--model_name_or_path internlm/internlm-7b \ | |
--dataset_name_or_path tatsu-lab/alpaca \ | |
# HuggingFace's default training args | |
--do_train = True | |
--per_device_train_batch_size = 1 | |
--learning_rate = 2e-5 | |
--save_strategy = 'epoch' | |
--lr_scheduler_type = 'cosine' | |
--logging_steps = 1 | |
``` | |