2021-07-05 00:58:54 +03:00
|
|
|
#! /bin/bash
|
|
|
|
./run_clm_flax.py \
|
|
|
|
--output_dir $HOME/gpt-neo-2.7B-code-clippy \
|
|
|
|
--model_name_or_path="EleutherAI/gpt-neo-2.7B" \
|
|
|
|
--dataset_name="code_search_net" \
|
|
|
|
--dataset_config_name="python" \
|
2021-07-05 21:24:02 +03:00
|
|
|
--text_column_name="func_code_string" \
|
2021-07-05 00:58:54 +03:00
|
|
|
--do_train --do_eval \
|
|
|
|
--block_size="128" \
|
|
|
|
--per_device_train_batch_size="1" \
|
|
|
|
--per_device_eval_batch_size="1" \
|
2021-07-05 21:24:02 +03:00
|
|
|
--preprocessing_num_workers="8" \
|
2021-07-05 00:58:54 +03:00
|
|
|
--dtype="bfloat16" \
|
|
|
|
--learning_rate="5e-3" \
|
|
|
|
--warmup_steps="1000" \
|
|
|
|
--adam_beta1="0.9" \
|
|
|
|
--adam_beta2="0.98" \
|
|
|
|
--weight_decay="0.01" \
|
|
|
|
--overwrite_output_dir \
|
|
|
|
--num_train_epochs="1" \
|
2021-07-08 14:34:16 +03:00
|
|
|
--push_to_hub="False" \
|
|
|
|
--resume_from_checkpoint="False"
|