Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- pt
|
4 |
+
---
|
5 |
+
|
6 |
+
``` python
|
7 |
+
|
8 |
+
from transformers import ElectraForPreTraining, ElectraTokenizerFast
|
9 |
+
import torch
|
10 |
+
|
11 |
+
discriminator = ElectraForPreTraining.from_pretrained("josu/electra-pt-br-small-discirminator")
|
12 |
+
tokenizer = ElectraTokenizerFast.from_pretrained("josu/electra-pt-br-small-discirminator")
|
13 |
+
|
14 |
+
sentence = "os passaros estão cantando"
|
15 |
+
fake_sentence = "os passaros estão falando"
|
16 |
+
|
17 |
+
fake_tokens = tokenizer.tokenize(fake_sentence)
|
18 |
+
fake_inputs = tokenizer.encode(fake_sentence, return_tensors="pt")
|
19 |
+
discriminator_outputs = discriminator(fake_inputs)
|
20 |
+
predictions = torch.round((torch.sign(discriminator_outputs[0]) + 1) / 2)
|
21 |
+
|
22 |
+
[print("%7s" % token, end="") for token in fake_tokens]
|
23 |
+
|
24 |
+
[print("%7s" % int(prediction), end="") for prediction in predictions.squeeze().tolist()
|
25 |
+
|
26 |
+
```
|