When I use the eval.py script for evaluting the model, the following command lines would be modified:
python eval.py \
--model_name_or_path t5-base-qg-hl \
--valid_file_path valid_data_qg_hl_t5.pt \
--model_type t5 \
--num_beams 4 \
--max_decoding_length 32 \
--output_path hypothesis_t5-base-qg-hl.txt
I replace the word "base" with "small" twice, then the command lines can run correctly. The following is the modified results:
python eval.py \
--model_name_or_path t5-small-qg-hl \
--valid_file_path valid_data_qg_hl_t5.pt \
--model_type t5 \
--num_beams 4 \
--max_decoding_length 32 \
--output_path hypothesis_t5-small-qg-hl.txt
When I use the eval.py script for evaluting the model, the following command lines would be modified: python eval.py \ --model_name_or_path t5-base-qg-hl \ --valid_file_path valid_data_qg_hl_t5.pt \ --model_type t5 \ --num_beams 4 \ --max_decoding_length 32 \ --output_path hypothesis_t5-base-qg-hl.txt
I replace the word "base" with "small" twice, then the command lines can run correctly. The following is the modified results: python eval.py \ --model_name_or_path t5-small-qg-hl \ --valid_file_path valid_data_qg_hl_t5.pt \ --model_type t5 \ --num_beams 4 \ --max_decoding_length 32 \ --output_path hypothesis_t5-small-qg-hl.txt