run_test.sh

jbsub -queue x86_1h -cores 4+1 -mem 30g -require a100 -o outputs/train/tweet_eval2/sentiment/deberta-v3-large-sentiment-lr5e-6-gas2-ls0.0/test.log /dccstor/tslm/envs/anaconda3/envs/tslm-gen/bin/python train_clf.py --model_name_or_path outputs/train/tweet_eval2/sentiment/deberta-v3-large-sentiment-lr5e-6-gas2-ls0.0/best_checkpoint --train_file data/tweet_eval/sentiment/train.csv --validation_file data/tweet_eval/sentiment/validation.csv --test_file data/tweet_eval/sentiment/test.csv --do_eval --do_predict --report_to none --per_device_eval_batch_size 16 --max_seq_length 256 --output_dir outputs/train/tweet_eval2/sentiment/deberta-v3-large-sentiment-lr5e-6-gas2-ls0.0/best_checkpoint