#!/bin/bash
#SBATCH --job-name=eval_cascaded_SLU_NLU.mt5-base.task_type-1.fine_tune.gpu_a100-40g+.node-1x1.bsz-64.epochs-22.metric-ema.metric_lang-all/checkpoint-30407
#SBATCH -n 1
#SBATCH -N 1
#SBATCH -p gpu
#SBATCH --gres=gpu:1
#SBATCH --cpus-per-task=8
#SBATCH --constraint=gpu_a100&gpu_40g
#SBATCH --mem=32G
#SBATCH --mail-type=ALL
#SBATCH --mail-user=beomseok.lee-phdstudent@naverlabs.com
#SBATCH --output=/beegfs/scratch/user/blee/project_3/models/NLU.mt5-base.task_type-1.fine_tune.gpu_a100-40g+.node-1x1.bsz-64.epochs-22.metric-ema.metric_lang-all/checkpoint-30407/sbatch-%j-02-05.22-53.log

source /home/blee/environments/py39-hugging-face/bin/activate

export http_proxy=http://proxy.int.europe.naverlabs.com:3128
export https_proxy=http://proxy.int.europe.naverlabs.com:3128
export no_proxy=int.europe.naverlabs.com

export HF_HOME=/beegfs/scratch/user/blee/hugging-face/models
export HF_DATASETS_DOWNLOADED_DATASETS_PATH=/beegfs/scratch/user/blee/hugging-face/downloaded
export HF_DATASETS_EXTRACTED_DATASETS_PATH=/beegfs/scratch/user/blee/hugging-face/extracted

export PYTHONPATH=:/home/blee/code-repo/transformers-slu

python /home/blee/code-repo/transformers-slu/nle/examples/nlu/run_nlu_mT5.py \
	--do_predict \
	--predict_with_generate \
	--use_fast_tokenizer \
	--trust_remote_code \
	--test_dataset_name /home/blee/code-repo/transformers-slu/nle/dataset/speech_massive_cascaded \
	--test_dataset_config_name multilingual-test \
	--model_name_or_path /beegfs/scratch/user/blee/project_3/models/NLU.mt5-base.task_type-1.fine_tune.gpu_a100-40g+.node-1x1.bsz-64.epochs-22.metric-ema.metric_lang-all/checkpoint-30407 \
	--output_dir /beegfs/scratch/user/blee/project_3/models/NLU.mt5-base.task_type-1.fine_tune.gpu_a100-40g+.node-1x1.bsz-64.epochs-22.metric-ema.metric_lang-all/checkpoint-30407/eval/cascaded_SLU \
	--preprocessing_num_workers 1 \
	--length_column_name input_length \
	--per_device_eval_batch_size 32 \
	--group_by_length \
	--overwrite_cache \
	--generation_num_beams 2