commit files to HF hub
Browse files
README.md
CHANGED
@@ -31,33 +31,33 @@ model-index:
|
|
31 |
metrics:
|
32 |
- name: BLEU4 (Question Answering)
|
33 |
type: bleu4_question_answering
|
34 |
-
value:
|
35 |
- name: ROUGE-L (Question Answering)
|
36 |
type: rouge_l_question_answering
|
37 |
-
value:
|
38 |
- name: METEOR (Question Answering)
|
39 |
type: meteor_question_answering
|
40 |
-
value:
|
41 |
- name: BERTScore (Question Answering)
|
42 |
type: bertscore_question_answering
|
43 |
-
value:
|
44 |
- name: MoverScore (Question Answering)
|
45 |
type: moverscore_question_answering
|
46 |
-
value:
|
47 |
- name: AnswerF1Score (Question Answering)
|
48 |
type: answer_f1_score__question_answering
|
49 |
-
value:
|
50 |
- name: AnswerExactMatch (Question Answering)
|
51 |
type: answer_exact_match_question_answering
|
52 |
-
value:
|
53 |
---
|
54 |
|
55 |
# Model Card of `vocabtrimmer/mt5-small-trimmed-fr-15000-frquad-qa`
|
56 |
-
This model is fine-tuned version of [
|
57 |
|
58 |
|
59 |
### Overview
|
60 |
-
- **Language model:** [
|
61 |
- **Language:** fr
|
62 |
- **Training data:** [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) (default)
|
63 |
- **Online Demo:** [https://autoqg.net/](https://autoqg.net/)
|
@@ -93,16 +93,16 @@ output = pipe("question: En quelle année a-t-on trouvé trace d'un haut fournea
|
|
93 |
|
94 |
| | Score | Type | Dataset |
|
95 |
|:-----------------|--------:|:--------|:-----------------------------------------------------------------|
|
96 |
-
| AnswerExactMatch |
|
97 |
-
| AnswerF1Score |
|
98 |
-
| BERTScore |
|
99 |
-
| Bleu_1 |
|
100 |
-
| Bleu_2 |
|
101 |
-
| Bleu_3 |
|
102 |
-
| Bleu_4 |
|
103 |
-
| METEOR |
|
104 |
-
| MoverScore |
|
105 |
-
| ROUGE_L |
|
106 |
|
107 |
|
108 |
|
@@ -114,12 +114,12 @@ The following hyperparameters were used during fine-tuning:
|
|
114 |
- input_types: ['paragraph_question']
|
115 |
- output_types: ['answer']
|
116 |
- prefix_types: None
|
117 |
-
- model:
|
118 |
- max_length: 512
|
119 |
- max_length_output: 32
|
120 |
-
- epoch:
|
121 |
- batch: 32
|
122 |
-
- lr: 0.
|
123 |
- fp16: False
|
124 |
- random_seed: 1
|
125 |
- gradient_accumulation_steps: 4
|
|
|
31 |
metrics:
|
32 |
- name: BLEU4 (Question Answering)
|
33 |
type: bleu4_question_answering
|
34 |
+
value: 17.12
|
35 |
- name: ROUGE-L (Question Answering)
|
36 |
type: rouge_l_question_answering
|
37 |
+
value: 27.44
|
38 |
- name: METEOR (Question Answering)
|
39 |
type: meteor_question_answering
|
40 |
+
value: 21.75
|
41 |
- name: BERTScore (Question Answering)
|
42 |
type: bertscore_question_answering
|
43 |
+
value: 88.56
|
44 |
- name: MoverScore (Question Answering)
|
45 |
type: moverscore_question_answering
|
46 |
+
value: 70.15
|
47 |
- name: AnswerF1Score (Question Answering)
|
48 |
type: answer_f1_score__question_answering
|
49 |
+
value: 43.25
|
50 |
- name: AnswerExactMatch (Question Answering)
|
51 |
type: answer_exact_match_question_answering
|
52 |
+
value: 26.88
|
53 |
---
|
54 |
|
55 |
# Model Card of `vocabtrimmer/mt5-small-trimmed-fr-15000-frquad-qa`
|
56 |
+
This model is fine-tuned version of [ckpts/mt5-small-trimmed-fr-15000](https://huggingface.co/ckpts/mt5-small-trimmed-fr-15000) for question answering task on the [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) (dataset_name: default) via [`lmqg`](https://github.com/asahi417/lm-question-generation).
|
57 |
|
58 |
|
59 |
### Overview
|
60 |
+
- **Language model:** [ckpts/mt5-small-trimmed-fr-15000](https://huggingface.co/ckpts/mt5-small-trimmed-fr-15000)
|
61 |
- **Language:** fr
|
62 |
- **Training data:** [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) (default)
|
63 |
- **Online Demo:** [https://autoqg.net/](https://autoqg.net/)
|
|
|
93 |
|
94 |
| | Score | Type | Dataset |
|
95 |
|:-----------------|--------:|:--------|:-----------------------------------------------------------------|
|
96 |
+
| AnswerExactMatch | 26.88 | default | [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) |
|
97 |
+
| AnswerF1Score | 43.25 | default | [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) |
|
98 |
+
| BERTScore | 88.56 | default | [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) |
|
99 |
+
| Bleu_1 | 25.9 | default | [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) |
|
100 |
+
| Bleu_2 | 22.1 | default | [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) |
|
101 |
+
| Bleu_3 | 19.45 | default | [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) |
|
102 |
+
| Bleu_4 | 17.12 | default | [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) |
|
103 |
+
| METEOR | 21.75 | default | [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) |
|
104 |
+
| MoverScore | 70.15 | default | [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) |
|
105 |
+
| ROUGE_L | 27.44 | default | [lmqg/qg_frquad](https://huggingface.co/datasets/lmqg/qg_frquad) |
|
106 |
|
107 |
|
108 |
|
|
|
114 |
- input_types: ['paragraph_question']
|
115 |
- output_types: ['answer']
|
116 |
- prefix_types: None
|
117 |
+
- model: ckpts/mt5-small-trimmed-fr-15000
|
118 |
- max_length: 512
|
119 |
- max_length_output: 32
|
120 |
+
- epoch: 26
|
121 |
- batch: 32
|
122 |
+
- lr: 0.001
|
123 |
- fp16: False
|
124 |
- random_seed: 1
|
125 |
- gradient_accumulation_steps: 4
|
eval/metric.first.answer.paragraph_question.answer.lmqg_qg_frquad.default.json
CHANGED
@@ -1 +1 @@
|
|
1 |
-
{"validation": {"Bleu_1":
|
|
|
1 |
+
{"validation": {"Bleu_1": 0.2664155005381952, "Bleu_2": 0.22753591111355484, "Bleu_3": 0.1993707280269415, "Bleu_4": 0.17659101079670417, "METEOR": 0.20698995597958483, "ROUGE_L": 0.2601730968698664, "BERTScore": 0.8857114457687739, "MoverScore": 0.6889768436401756, "AnswerF1Score": 42.29288054020749, "AnswerExactMatch": 22.45922208281054}, "test": {"Bleu_1": 0.25904334828099707, "Bleu_2": 0.22102083058968738, "Bleu_3": 0.19446005005164355, "Bleu_4": 0.17115253649023965, "METEOR": 0.2174977961913064, "ROUGE_L": 0.2743645114245147, "BERTScore": 0.885604128110663, "MoverScore": 0.7015155928389406, "AnswerF1Score": 43.253076879987645, "AnswerExactMatch": 26.882057716436638}}
|
eval/samples.test.hyp.paragraph_question.answer.lmqg_qg_frquad.default.txt
CHANGED
The diff for this file is too large to render.
See raw diff
|
|
eval/samples.validation.hyp.paragraph_question.answer.lmqg_qg_frquad.default.txt
CHANGED
The diff for this file is too large to render.
See raw diff
|
|