export RUN_NAME=test

./venv/bin/python train.py \
--model_name_or_path="output/test/flax_model.msgpack" \
--t5_model_name_or_path="output/test/flax_model.msgpack" \
--output_dir="output/${RUN_NAME}" \
--overwrite_output_dir \
--dataset_name="ag_news" \
--do_train --do_eval \
--n_latent_tokens 6 \
--latent_token_size 32 \
--save_steps="2500" \
--eval_steps="2500" \
--block_size="32" \
--per_device_train_batch_size="1" \
--per_device_eval_batch_size="1" \
--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" \