Update README.md
Browse files
README.md
CHANGED
@@ -20,7 +20,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
20 |
|
21 |
# xlm-roberta-base-finetuned-JaQuAD
|
22 |
|
23 |
-
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the
|
24 |
It achieves the following results on the evaluation set:
|
25 |
- Loss: 0.7495
|
26 |
|
@@ -33,6 +33,24 @@ More information needed
|
|
33 |
```python
|
34 |
import torch
|
35 |
from transformers import AutoModelForQuestionAnswering, AutoTokenizer
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
36 |
```
|
37 |
|
38 |
## Training and evaluation data
|
|
|
20 |
|
21 |
# xlm-roberta-base-finetuned-JaQuAD
|
22 |
|
23 |
+
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the [JaQuAD](https://huggingface.co/datasets/SkelterLabsInc/JaQuAD) dataset.
|
24 |
It achieves the following results on the evaluation set:
|
25 |
- Loss: 0.7495
|
26 |
|
|
|
33 |
```python
|
34 |
import torch
|
35 |
from transformers import AutoModelForQuestionAnswering, AutoTokenizer
|
36 |
+
|
37 |
+
text = "2015年9月1日、私は横浜へ車で出かけました。映画を観た後に中華街まで電車で行き、昼ご飯は重慶飯店で中華フルコースを食べました。"
|
38 |
+
questions= ["どこへ出かけた?", "電車に乗る前は何をしていた?", "重慶飯店で何を食べた?", "いつ横浜に出かけた?"]
|
39 |
+
|
40 |
+
for question in questions:
|
41 |
+
|
42 |
+
inputs = tokenizer.encode_plus(question, text, add_special_tokens=True, return_tensors="pt")
|
43 |
+
|
44 |
+
with torch.no_grad():
|
45 |
+
output = model(**inputs)
|
46 |
+
|
47 |
+
answer_start = torch.argmax(output.start_logits)
|
48 |
+
answer_end = torch.argmax(output.end_logits)
|
49 |
+
|
50 |
+
answer_tokens = inputs.input_ids[0, answer_start : answer_end + 1]
|
51 |
+
answer = tokenizer.decode(answer_tokens)
|
52 |
+
|
53 |
+
print(f"質問: {question} -> 回答: {answer}")
|
54 |
```
|
55 |
|
56 |
## Training and evaluation data
|