asahi417 commited on
Commit
5a60e61
1 Parent(s): a0c264d

model update

Browse files
README.md ADDED
@@ -0,0 +1,215 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ ---
3
+ license: cc-by-4.0
4
+ metrics:
5
+ - bleu4
6
+ - meteor
7
+ - rouge-l
8
+ - bertscore
9
+ - moverscore
10
+ language: ko
11
+ datasets:
12
+ - lmqg/qg_koquad
13
+ pipeline_tag: text2text-generation
14
+ tags:
15
+ - question generation
16
+ - answer extraction
17
+ widget:
18
+ - text: "generate question: 1990년 영화 《 <hl> 남부군 <hl> 》에서 단역으로 영화배우 첫 데뷔에 이어 같은 해 KBS 드라마 《지구인》에서 단역으로 출연하였고 이듬해 MBC 《여명의 눈동자》를 통해 단역으로 출연하였다."
19
+ example_title: "Question Generation Example 1"
20
+ - text: "generate question: 백신이 없기때문에 예방책은 <hl> 살충제 <hl> 를 사용하면서 서식 장소(찻찬 받침, 배수로, 고인 물의 열린 저장소, 버려진 타이어 등)의 수를 줄임으로써 매개체를 통제할 수 있다."
21
+ example_title: "Question Generation Example 2"
22
+ - text: "generate question: <hl> 원테이크 촬영 <hl> 이기 때문에 한 사람이 실수를 하면 처음부터 다시 찍어야 하는 상황이 발생한다."
23
+ example_title: "Question Generation Example 3"
24
+ - text: "extract answers: 또한 스피어스는 많은 새로운 여성 아티스트들에게 영향을 끼쳤는데, 대표적으로 데미 로바토, 케이티 페리, 크리스티니아 드바지, 레이디 가가, 리틀 부츠, 셀레나 고메즈 & 더씬, 픽시 로트 이 있다. 2007년 비욘세 놀스는 Total Request Live와의 인터뷰에서 '나는 브리트니를 사랑하고 팬이에요. 특히 새 앨범 Blackout을 좋아해요'라고 말했다. 린제이 로한은 '언제나 브리트니 스피어스에게 영감을 받는다. 학창시절 그녀처럼 타블로이드에 오르기를 꿈꿔왔다'고 말하며 롤 모델로 꼽았다. 스피어스는 현대 음악가들에게 음악적 영감으로 언급되기도 했다. <hl> 마일리 사이러스는 자신의 히트곡 Party in the U.S.A. 가 브리트니에게 영감과 영향을 받은 곡이라고 밝혔다. <hl> 베리 매닐로우의 앨범 15 Minutes 역시 브리트니에게 영감을 얻었다고 언급되었다."
25
+ example_title: "Answer Extraction Example 1"
26
+ - text: "extract answers: 지난 22일 아프리카TV는 BJ 철구가 서비스 정지 처분을 받았음을 밝혔다. 서비스 정지 처분을 사유는 철구가 10대 청소년에게 유해한 장면을 방송으로 내보냈기 때문이었다. 문제가 된 장면은 BJ 철구가 미성년자는 시청할 수 없게 하는 19세 시청 가능 설정을 하지 않은 채 흡연하는 모습을 여과 없이 드러낸 장면이다. 아프리카TV는 청소년 보호 정책의 '청소년들이 해로운 환경으로부터 보호받을 수 있도록 조치한다'라고 조항을 근거로 철구에게 서비스 정지 처분을 내렸다. 흡연 이외에 음주 방송 등도 19세 시청 가능 설정을 해야만 방송할 수 있다. <hl> 게다가 철구의 방송 정지 처분은 이번에 처음이 아니라 16번 째기 때문에 더욱더 논란이 되고 있다. <hl>"
27
+ example_title: "Answer Extraction Example 2"
28
+ model-index:
29
+ - name: lmqg/mbart-large-cc25-koquad-qg-ae
30
+ results:
31
+ - task:
32
+ name: Text2text Generation
33
+ type: text2text-generation
34
+ dataset:
35
+ name: lmqg/qg_koquad
36
+ type: default
37
+ args: default
38
+ metrics:
39
+ - name: BLEU4 (Question Generation)
40
+ type: bleu4_question_generation
41
+ value: 10.7
42
+ - name: ROUGE-L (Question Generation)
43
+ type: rouge_l_question_generation
44
+ value: 27.02
45
+ - name: METEOR (Question Generation)
46
+ type: meteor_question_generation
47
+ value: 29.73
48
+ - name: BERTScore (Question Generation)
49
+ type: bertscore_question_generation
50
+ value: 83.52
51
+ - name: MoverScore (Question Generation)
52
+ type: moverscore_question_generation
53
+ value: 82.79
54
+ - name: QAAlignedF1Score-BERTScore (Question & Answer Generation (with Gold Answer))
55
+ type: qa_aligned_f1_score_bertscore_question_answer_generation_with_gold_answer
56
+ value: 80.81
57
+ - name: QAAlignedRecall-BERTScore (Question & Answer Generation (with Gold Answer))
58
+ type: qa_aligned_recall_bertscore_question_answer_generation_with_gold_answer
59
+ value: 84.32
60
+ - name: QAAlignedPrecision-BERTScore (Question & Answer Generation (with Gold Answer))
61
+ type: qa_aligned_precision_bertscore_question_answer_generation_with_gold_answer
62
+ value: 77.64
63
+ - name: QAAlignedF1Score-MoverScore (Question & Answer Generation (with Gold Answer))
64
+ type: qa_aligned_f1_score_moverscore_question_answer_generation_with_gold_answer
65
+ value: 83.42
66
+ - name: QAAlignedRecall-MoverScore (Question & Answer Generation (with Gold Answer))
67
+ type: qa_aligned_recall_moverscore_question_answer_generation_with_gold_answer
68
+ value: 88.44
69
+ - name: QAAlignedPrecision-MoverScore (Question & Answer Generation (with Gold Answer))
70
+ type: qa_aligned_precision_moverscore_question_answer_generation_with_gold_answer
71
+ value: 79.08
72
+ - name: BLEU4 (Answer Extraction)
73
+ type: bleu4_answer_extraction
74
+ value: 24.34
75
+ - name: ROUGE-L (Answer Extraction)
76
+ type: rouge_l_answer_extraction
77
+ value: 82.78
78
+ - name: METEOR (Answer Extraction)
79
+ type: meteor_answer_extraction
80
+ value: 59.82
81
+ - name: BERTScore (Answer Extraction)
82
+ type: bertscore_answer_extraction
83
+ value: 95.53
84
+ - name: MoverScore (Answer Extraction)
85
+ type: moverscore_answer_extraction
86
+ value: 94.69
87
+ - name: AnswerF1Score (Answer Extraction)
88
+ type: answer_f1_score__answer_extraction
89
+ value: 88.2
90
+ - name: AnswerExactMatch (Answer Extraction)
91
+ type: answer_exact_match_answer_extraction
92
+ value: 82.17
93
+ ---
94
+
95
+ # Model Card of `lmqg/mbart-large-cc25-koquad-qg-ae`
96
+ This model is fine-tuned version of [facebook/mbart-large-cc25](https://huggingface.co/facebook/mbart-large-cc25) for question generation and answer extraction jointly on the [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) (dataset_name: default) via [`lmqg`](https://github.com/asahi417/lm-question-generation).
97
+
98
+
99
+ ### Overview
100
+ - **Language model:** [facebook/mbart-large-cc25](https://huggingface.co/facebook/mbart-large-cc25)
101
+ - **Language:** ko
102
+ - **Training data:** [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) (default)
103
+ - **Online Demo:** [https://autoqg.net/](https://autoqg.net/)
104
+ - **Repository:** [https://github.com/asahi417/lm-question-generation](https://github.com/asahi417/lm-question-generation)
105
+ - **Paper:** [https://arxiv.org/abs/2210.03992](https://arxiv.org/abs/2210.03992)
106
+
107
+ ### Usage
108
+ - With [`lmqg`](https://github.com/asahi417/lm-question-generation#lmqg-language-model-for-question-generation-)
109
+ ```python
110
+ from lmqg import TransformersQG
111
+
112
+ # initialize model
113
+ model = TransformersQG(language="ko", model="lmqg/mbart-large-cc25-koquad-qg-ae")
114
+
115
+ # model prediction
116
+ question_answer_pairs = model.generate_qa("1990년 영화 《 남부군 》에서 단역으로 영화배우 첫 데뷔에 이어 같은 해 KBS 드라마 《지구인》에서 단역으로 출연하였고 이듬해 MBC 《여명의 눈동자》를 통해 단역으로 출연하였다.")
117
+
118
+ ```
119
+
120
+ - With `transformers`
121
+ ```python
122
+ from transformers import pipeline
123
+
124
+ pipe = pipeline("text2text-generation", "lmqg/mbart-large-cc25-koquad-qg-ae")
125
+
126
+ # answer extraction
127
+ answer = pipe("generate question: 1990년 영화 《 <hl> 남부군 <hl> 》에서 단역으로 영화배우 첫 데뷔에 이어 같은 해 KBS 드라마 《지구인》에서 단역으로 출연하였고 이듬해 MBC 《여명의 눈동자》를 통해 단역으로 출연하였다.")
128
+
129
+ # question generation
130
+ question = pipe("extract answers: 또한 스피어스는 많은 새로운 여성 아티스트들에게 영향을 끼쳤는데, 대표적으로 데미 로바토, 케이티 페리, 크리스티니아 드바지, 레이디 가가, 리틀 부츠, 셀레나 고메즈 & 더씬, 픽시 로트 이 있다. 2007년 비욘세 놀스는 Total Request Live와의 인터뷰에서 '나는 브리트니를 사랑하고 팬이에요. 특히 새 앨범 Blackout을 좋아해요'라고 말했다. 린제이 로한은 '언제나 브리트니 스피어스에게 영감을 받는다. 학창시절 그녀처럼 타블로이드에 오르기를 꿈꿔왔다'고 말하며 롤 모델로 꼽았다. 스피어스는 현대 음악가들에게 음악적 영감으로 언급되기도 했다. <hl> 마일리 사이러스는 자신의 히트곡 Party in the U.S.A. 가 브리트니에게 영감과 영향을 받은 곡이라고 밝혔다. <hl> 베리 매닐로우의 앨범 15 Minutes 역시 브리트니에게 영감을 얻었다고 언급되었다.")
131
+
132
+ ```
133
+
134
+ ## Evaluation
135
+
136
+
137
+ - ***Metric (Question Generation)***: [raw metric file](https://huggingface.co/lmqg/mbart-large-cc25-koquad-qg-ae/raw/main/eval/metric.first.sentence.paragraph_answer.question.lmqg_qg_koquad.default.json)
138
+
139
+ | | Score | Type | Dataset |
140
+ |:-----------|--------:|:--------|:-----------------------------------------------------------------|
141
+ | BERTScore | 83.52 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
142
+ | Bleu_1 | 26.03 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
143
+ | Bleu_2 | 18.93 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
144
+ | Bleu_3 | 14.14 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
145
+ | Bleu_4 | 10.7 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
146
+ | METEOR | 29.73 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
147
+ | MoverScore | 82.79 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
148
+ | ROUGE_L | 27.02 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
149
+
150
+
151
+ - ***Metric (Question & Answer Generation)***: [raw metric file](https://huggingface.co/lmqg/mbart-large-cc25-koquad-qg-ae/raw/main/eval/metric.first.answer.paragraph.questions_answers.lmqg_qg_koquad.default.json)
152
+
153
+ | | Score | Type | Dataset |
154
+ |:--------------------------------|--------:|:--------|:-----------------------------------------------------------------|
155
+ | QAAlignedF1Score (BERTScore) | 80.81 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
156
+ | QAAlignedF1Score (MoverScore) | 83.42 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
157
+ | QAAlignedPrecision (BERTScore) | 77.64 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
158
+ | QAAlignedPrecision (MoverScore) | 79.08 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
159
+ | QAAlignedRecall (BERTScore) | 84.32 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
160
+ | QAAlignedRecall (MoverScore) | 88.44 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
161
+
162
+
163
+ - ***Metric (Answer Extraction)***: [raw metric file](https://huggingface.co/lmqg/mbart-large-cc25-koquad-qg-ae/raw/main/eval/metric.first.answer.paragraph_sentence.answer.lmqg_qg_koquad.default.json)
164
+
165
+ | | Score | Type | Dataset |
166
+ |:-----------------|--------:|:--------|:-----------------------------------------------------------------|
167
+ | AnswerExactMatch | 82.17 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
168
+ | AnswerF1Score | 88.2 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
169
+ | BERTScore | 95.53 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
170
+ | Bleu_1 | 68.81 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
171
+ | Bleu_2 | 56.84 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
172
+ | Bleu_3 | 40.49 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
173
+ | Bleu_4 | 24.34 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
174
+ | METEOR | 59.82 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
175
+ | MoverScore | 94.69 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
176
+ | ROUGE_L | 82.78 | default | [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) |
177
+
178
+
179
+
180
+ ## Training hyperparameters
181
+
182
+ The following hyperparameters were used during fine-tuning:
183
+ - dataset_path: lmqg/qg_koquad
184
+ - dataset_name: default
185
+ - input_types: ['paragraph_answer', 'paragraph_sentence']
186
+ - output_types: ['question', 'answer']
187
+ - prefix_types: ['qg', 'ae']
188
+ - model: facebook/mbart-large-cc25
189
+ - max_length: 512
190
+ - max_length_output: 32
191
+ - epoch: 6
192
+ - batch: 2
193
+ - lr: 0.0001
194
+ - fp16: False
195
+ - random_seed: 1
196
+ - gradient_accumulation_steps: 32
197
+ - label_smoothing: 0.15
198
+
199
+ The full configuration can be found at [fine-tuning config file](https://huggingface.co/lmqg/mbart-large-cc25-koquad-qg-ae/raw/main/trainer_config.json).
200
+
201
+ ## Citation
202
+ ```
203
+ @inproceedings{ushio-etal-2022-generative,
204
+ title = "{G}enerative {L}anguage {M}odels for {P}aragraph-{L}evel {Q}uestion {G}eneration",
205
+ author = "Ushio, Asahi and
206
+ Alva-Manchego, Fernando and
207
+ Camacho-Collados, Jose",
208
+ booktitle = "Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing",
209
+ month = dec,
210
+ year = "2022",
211
+ address = "Abu Dhabi, U.A.E.",
212
+ publisher = "Association for Computational Linguistics",
213
+ }
214
+
215
+ ```
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "lmqg_output/mbart-large-cc25-koquad-qg-ae/best_model",
3
  "_num_labels": 3,
4
  "activation_dropout": 0.0,
5
  "activation_function": "gelu",
 
1
  {
2
+ "_name_or_path": "lmqg_output/mbart-large-cc25-koquad-qg-ae/model_yyyubh/epoch_5",
3
  "_num_labels": 3,
4
  "activation_dropout": 0.0,
5
  "activation_function": "gelu",
eval/metric.first.answer.paragraph.questions_answers.lmqg_qg_koquad.default.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"test": {"QAAlignedF1Score (BERTScore)": 0.8080531331560218, "QAAlignedRecall (BERTScore)": 0.8432085347440835, "QAAlignedPrecision (BERTScore)": 0.7764423377788053, "QAAlignedF1Score (MoverScore)": 0.8341647325006587, "QAAlignedRecall (MoverScore)": 0.884428816496664, "QAAlignedPrecision (MoverScore)": 0.790849078156105, "Bleu_1": 0.06006353283216381, "Bleu_2": 0.03277271597776233, "Bleu_3": 0.01679141278827258, "Bleu_4": 0.008701219699473002, "METEOR": 0.2096936779895271, "ROUGE_L": 0.10486478215766001, "BERTScore": 0.6530535750859281, "MoverScore": 0.6212690775930128}, "validation": {"QAAlignedF1Score (BERTScore)": 0.8320523455339407, "QAAlignedRecall (BERTScore)": 0.842912377202491, "QAAlignedPrecision (BERTScore)": 0.8218751313450386, "QAAlignedF1Score (MoverScore)": 0.8776802710910037, "QAAlignedRecall (MoverScore)": 0.8900046908054511, "QAAlignedPrecision (MoverScore)": 0.8670147916474056, "Bleu_1": 0.24769373372825015, "Bleu_2": 0.16276359134020726, "Bleu_3": 0.10296341046122665, "Bleu_4": 0.0631750284426749, "METEOR": 0.3244083271656011, "ROUGE_L": 0.2514414206314093, "BERTScore": 0.7588320937007665, "MoverScore": 0.6899094830883825}}
eval/metric.first.answer.paragraph_answer.question.lmqg_qg_koquad.default.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"validation": {"Bleu_1": 0.24001131861912395, "Bleu_2": 0.17337773126724246, "Bleu_3": 0.12863453775221245, "Bleu_4": 0.0967813335443571}, "test": {"Bleu_1": 0.25700589970500975, "Bleu_2": 0.18655624690855915, "Bleu_3": 0.13921999044020586, "Bleu_4": 0.10522212620417044}}
eval/metric.first.answer.paragraph_sentence.answer.lmqg_qg_koquad.default.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"validation": {"Bleu_1": 0.6473646468267616, "Bleu_2": 0.5223649094151351, "Bleu_3": 0.3708659604524931, "Bleu_4": 0.22928480043252533, "METEOR": 0.5810308270805657, "ROUGE_L": 0.8033567513331389, "BERTScore": 0.9471528558548317, "MoverScore": 0.9376238552757546, "AnswerF1Score": 85.70134600751668, "AnswerExactMatch": 79.62192160943462}, "test": {"Bleu_1": 0.6880666727721688, "Bleu_2": 0.5684242128732656, "Bleu_3": 0.4048775421649216, "Bleu_4": 0.24343180251763877, "METEOR": 0.5982393486155895, "ROUGE_L": 0.8277957798666822, "BERTScore": 0.9553042152173561, "MoverScore": 0.9469417855096769, "AnswerF1Score": 88.20101741568419, "AnswerExactMatch": 82.17134928893513}}
eval/metric.first.sentence.paragraph_answer.question.lmqg_qg_koquad.default.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"validation": {"Bleu_1": 0.27035303490513696, "Bleu_2": 0.19930839007285978, "Bleu_3": 0.1501884183509421, "Bleu_4": 0.11424533846569236, "METEOR": 0.30116603338636466, "ROUGE_L": 0.2760626109540457, "BERTScore": 0.8260049123941742, "MoverScore": 0.8266965044904961}, "test": {"Bleu_1": 0.2602554141015628, "Bleu_2": 0.18932977159555883, "Bleu_3": 0.1414435531316931, "Bleu_4": 0.10703670501766625, "METEOR": 0.2973001847077773, "ROUGE_L": 0.27015462949790164, "BERTScore": 0.83523548089995, "MoverScore": 0.8278969743029401}}
eval/samples.test.hyp.paragraph.questions_answers.lmqg_qg_koquad.default.txt ADDED
The diff for this file is too large to render. See raw diff
 
eval/samples.test.hyp.paragraph_answer.question.lmqg_qg_koquad.default.txt ADDED
The diff for this file is too large to render. See raw diff
 
eval/samples.test.hyp.paragraph_sentence.answer.lmqg_qg_koquad.default.txt ADDED
The diff for this file is too large to render. See raw diff
 
eval/samples.validation.hyp.paragraph.questions_answers.lmqg_qg_koquad.default.txt ADDED
The diff for this file is too large to render. See raw diff
 
eval/samples.validation.hyp.paragraph_answer.question.lmqg_qg_koquad.default.txt ADDED
The diff for this file is too large to render. See raw diff
 
eval/samples.validation.hyp.paragraph_sentence.answer.lmqg_qg_koquad.default.txt ADDED
The diff for this file is too large to render. See raw diff
 
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5d4dc071c41971fde00640693667717d0f5a277d6667d6abded65f96fff93650
3
- size 2444580125
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c7130820f5d338a88c9eb376f13f754fff44599066d6a537029a720bd1f12364
3
+ size 2444587421
tokenizer_config.json CHANGED
@@ -12,7 +12,7 @@
12
  "single_word": false
13
  },
14
  "model_max_length": 1024,
15
- "name_or_path": "lmqg_output/mbart-large-cc25-koquad-qg-ae/best_model",
16
  "pad_token": "<pad>",
17
  "sep_token": "</s>",
18
  "special_tokens_map_file": null,
 
12
  "single_word": false
13
  },
14
  "model_max_length": 1024,
15
+ "name_or_path": "lmqg_output/mbart-large-cc25-koquad-qg-ae/model_yyyubh/epoch_5",
16
  "pad_token": "<pad>",
17
  "sep_token": "</s>",
18
  "special_tokens_map_file": null,
trainer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"dataset_path": "lmqg/qg_koquad", "dataset_name": "default", "input_types": ["paragraph_answer", "paragraph_sentence"], "output_types": ["question", "answer"], "prefix_types": ["qg", "ae"], "model": "facebook/mbart-large-cc25", "max_length": 512, "max_length_output": 32, "epoch": 6, "batch": 2, "lr": 0.0001, "fp16": false, "random_seed": 1, "gradient_accumulation_steps": 32, "label_smoothing": 0.15}