mirror of https://github.com/tensorflow/models.git
README.md
Code for Roformer.
Run with
DATA_PATH=???
OUTPUT_DIR=???
python3 train.py \
--experiment=roformer/pretraining \
--config_file=experiments/roformer_base.yaml \
--params_override="task.validation_data.input_path=${DATA_PATH},runtime.distribution_strategy=tpu" \
--tpu=local \
--model_dir=${OUTPUT_DIR} \
--mode=train_and_eval