transformers/examples/seq2seq/train_mbart_cc25_enro.sh
Sam Shleifer 57eb1cb68d
[s2s] Document better mbart finetuning command (#6229)
* Document better MT command

* improve multigpu command
2020-08-03 18:22:31 -04:00

19 lines
596 B
Bash
Executable File

#!/usr/bin/env bash
export PYTHONPATH="../":"${PYTHONPATH}"
python finetune.py \
--learning_rate=3e-5 \
--fp16 \
--do_train \
--val_check_interval=0.25 \
--adam_eps 1e-06 \
--num_train_epochs 6 --src_lang en_XX --tgt_lang ro_RO \
--data_dir $ENRO_DIR \
--max_source_length $MAX_LEN --max_target_length $MAX_LEN --val_max_target_length $MAX_LEN --test_max_target_length $MAX_LEN \
--train_batch_size=$BS --eval_batch_size=$BS \
--task translation \
--warmup_steps 500 \
--freeze_embeds \
--model_name_or_path=facebook/mbart-large-cc25 \
$@