change corpora and datasets

dbiir · Oct 1, 2023 · 9e1e635 · 9e1e635
1 parent 0efe195
commit 9e1e635
Show file tree

Hide file tree

Showing 16 changed files with 6 additions and 6 deletions.
diff --git a/.github/workflows/github-actions.yml b/.github/workflows/github-actions.yml
@@ -47,9 +47,9 @@ jobs:
               mv models/pegasus_model.bin-10 models/pegasus_model.bin
               python finetune/run_classifier.py --pretrained_model_path models/bert_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/bert/mini_config.json --output_model_path models/classifier_model.bin --train_path datasets/test_data/book_review/train.tsv --dev_path datasets/test_data/book_review/dev.tsv --epochs_num 3 --batch_size 2
               python inference/run_classifier_infer.py --load_model_path models/classifier_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/bert/mini_config.json --test_path datasets/test_data/book_review/test_nolabel.tsv --prediction_path datasets/test_data/book_review/prediction.tsv --labels_num 2
-              python finetune/run_classifier.py --pretrained_model_path models/albert_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/albert/base_config.json --output_model_path models/classifier_model.bin --train_path datasets/test_data/chnsenticorp_test/train.tsv --dev_path datasets/test_data/chnsenticorp_test/dev.tsv --learning_rate 4e-5 --epochs_num 3 --batch_size 2
-              python finetune/run_classifier_mt.py --pretrained_model_path models/bert_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/bert/mini_config.json --dataset_path_list datasets/test_data/douban_test/ datasets/test_data/chnsenticorp_test/ --epochs_num 1 --batch_size 2
-              python finetune/run_ner.py --pretrained_model_path models/bert_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/bert/mini_config.json --output_model_path models/ner_model.bin --train_path datasets/test_data/msra_ner_test/train.tsv --dev_path datasets/test_data/msra_ner_test/dev.tsv --label2id_path datasets/msra_ner/label2id.json --epochs_num 2 --batch_size 2
-              python inference/run_ner_infer.py --load_model_path models/ner_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/bert/mini_config.json --test_path datasets/test_data/msra_ner_test/test_nolabel.tsv --prediction_path datasets/test_data/msra_ner_test/prediction.tsv --label2id_path datasets/msra_ner/label2id.json
-              python finetune/run_cmrc.py --pretrained_model_path models/bert_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/bert/mini_config.json --output_model_path models/cmrc_model.bin --train_path datasets/test_data/cmrc_test/train.json --dev_path datasets/test_data/cmrc_test/dev.json --epochs_num 2 --batch_size 2 --seq_length 128
-              python inference/run_cmrc_infer.py --load_model_path models/cmrc_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/bert/mini_config.json --test_path datasets/test_data/cmrc_test/test.json --prediction_path datasets/test_data/cmrc_test/prediction.json --seq_length 128
+              python finetune/run_classifier.py --pretrained_model_path models/albert_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/albert/base_config.json --output_model_path models/classifier_model.bin --train_path datasets/test_data/chnsenticorp/train.tsv --dev_path datasets/test_data/chnsenticorp/dev.tsv --learning_rate 4e-5 --epochs_num 3 --batch_size 2
+              python finetune/run_classifier_mt.py --pretrained_model_path models/bert_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/bert/mini_config.json --dataset_path_list datasets/test_data/book_review/ datasets/test_data/chnsenticorp/ --epochs_num 1 --batch_size 2
+              python finetune/run_ner.py --pretrained_model_path models/bert_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/bert/mini_config.json --output_model_path models/ner_model.bin --train_path datasets/test_data/msra_ner/train.tsv --dev_path datasets/test_data/msra_ner/dev.tsv --label2id_path datasets/msra_ner/label2id.json --epochs_num 2 --batch_size 2
+              python inference/run_ner_infer.py --load_model_path models/ner_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/bert/mini_config.json --test_path datasets/test_data/msra_ner/test_nolabel.tsv --prediction_path datasets/test_data/msra_ner/prediction.tsv --label2id_path datasets/msra_ner/label2id.json
+              python finetune/run_cmrc.py --pretrained_model_path models/bert_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/bert/mini_config.json --output_model_path models/cmrc_model.bin --train_path datasets/test_data/cmrc/train.json --dev_path datasets/test_data/cmrc/dev.json --epochs_num 2 --batch_size 2 --seq_length 128
+              python inference/run_cmrc_infer.py --load_model_path models/cmrc_model.bin --vocab_path models/google_zh_vocab.txt --config_path models/bert/mini_config.json --test_path datasets/test_data/cmrc/test.json --prediction_path datasets/test_data/cmrc/prediction.json --seq_length 128
diff --git a/datasets/test_data/douban_test/dev.tsv → datasets/test_data/book_review/dev.tsv b/datasets/test_data/douban_test/dev.tsv → datasets/test_data/book_review/dev.tsv
diff --git a/datasets/test_data/douban_test/test.tsv → datasets/test_data/book_review/test.tsv b/datasets/test_data/douban_test/test.tsv → datasets/test_data/book_review/test.tsv
diff --git a/...ts/test_data/douban_test/test_nolabel.tsv → ...ts/test_data/book_review/test_nolabel.tsv b/...ts/test_data/douban_test/test_nolabel.tsv → ...ts/test_data/book_review/test_nolabel.tsv
diff --git a/datasets/test_data/douban_test/train.tsv → datasets/test_data/book_review/train.tsv b/datasets/test_data/douban_test/train.tsv → datasets/test_data/book_review/train.tsv
diff --git a/datasets/test_data/chnsenticorp_test/dev.tsv → datasets/test_data/chnsenticorp/dev.tsv b/datasets/test_data/chnsenticorp_test/dev.tsv → datasets/test_data/chnsenticorp/dev.tsv
diff --git a/...sets/test_data/chnsenticorp_test/test.tsv → datasets/test_data/chnsenticorp/test.tsv b/...sets/test_data/chnsenticorp_test/test.tsv → datasets/test_data/chnsenticorp/test.tsv
diff --git a/...t_data/chnsenticorp_test/test_nolabel.tsv → ...s/test_data/chnsenticorp/test_nolabel.tsv b/...t_data/chnsenticorp_test/test_nolabel.tsv → ...s/test_data/chnsenticorp/test_nolabel.tsv
diff --git a/...ets/test_data/chnsenticorp_test/train.tsv → datasets/test_data/chnsenticorp/train.tsv b/...ets/test_data/chnsenticorp_test/train.tsv → datasets/test_data/chnsenticorp/train.tsv
diff --git a/datasets/test_data/cmrc_test/dev.json → datasets/test_data/cmrc/dev.json b/datasets/test_data/cmrc_test/dev.json → datasets/test_data/cmrc/dev.json
diff --git a/datasets/test_data/cmrc_test/test.json → datasets/test_data/cmrc/test.json b/datasets/test_data/cmrc_test/test.json → datasets/test_data/cmrc/test.json
diff --git a/datasets/test_data/cmrc_test/train.json → datasets/test_data/cmrc/train.json b/datasets/test_data/cmrc_test/train.json → datasets/test_data/cmrc/train.json
diff --git a/datasets/test_data/msra_ner_test/dev.tsv → datasets/test_data/msra_ner/dev.tsv b/datasets/test_data/msra_ner_test/dev.tsv → datasets/test_data/msra_ner/dev.tsv
diff --git a/datasets/test_data/msra_ner_test/test.tsv → datasets/test_data/msra_ner/test.tsv b/datasets/test_data/msra_ner_test/test.tsv → datasets/test_data/msra_ner/test.tsv
diff --git a/.../test_data/msra_ner_test/test_nolabel.tsv → datasets/test_data/msra_ner/test_nolabel.tsv b/.../test_data/msra_ner_test/test_nolabel.tsv → datasets/test_data/msra_ner/test_nolabel.tsv
diff --git a/datasets/test_data/msra_ner_test/train.tsv → datasets/test_data/msra_ner/train.tsv b/datasets/test_data/msra_ner_test/train.tsv → datasets/test_data/msra_ner/train.tsv