Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,24 @@
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
---
|
| 4 |
+
|
| 5 |
+
## Usage
|
| 6 |
+
```
|
| 7 |
+
def qa(doc, q):
|
| 8 |
+
doc = doc.replace('\n',' ')
|
| 9 |
+
q = q.replace('\n',' ')
|
| 10 |
+
q_pr = f'<SC6>Опираясь на информацию: {doc}\n ответь на вопрос: \"{q}\".\n Ответ: '
|
| 11 |
+
data_inp = tokenizer(q_pr, return_tensors="pt").to('cuda:0')
|
| 12 |
+
return data_inp
|
| 13 |
+
|
| 14 |
+
def generate(doc, q):
|
| 15 |
+
t = qa(doc, q)
|
| 16 |
+
output_ids = model.generate(
|
| 17 |
+
**t, do_sample=False, temperature=0.0, max_new_tokens=512, repetition_penalty=1, no_repeat_ngram_size=8
|
| 18 |
+
)[0]
|
| 19 |
+
out = tokenizer.decode(output_ids.tolist(), skip_special_tokens=True)
|
| 20 |
+
out = out.replace("<extra_id_0>","")
|
| 21 |
+
ans_sqs = sent_tokenize(out, language="russian")
|
| 22 |
+
ans = ' '.join(ans_sqs[:3])
|
| 23 |
+
return ans.split('Ответ:')[0].split('Вопрос:')[0]
|
| 24 |
+
```
|