Update README.md
Browse files
README.md
CHANGED
|
@@ -1,34 +1,33 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
---
|
| 4 |
-
<h2>Re-Punctuate:</h2>
|
| 5 |
-
|
| 6 |
-
Re-Punctuate is a T5 model that attempts to correct Capitalization and Punctuations in the sentences.
|
| 7 |
-
|
| 8 |
-
<h3>DataSet:</h3>
|
| 9 |
-
|
| 10 |
-
DialogSum dataset (115056 Records) was used to fine-tune the model for Punctuation and Capitalization correction.
|
| 11 |
-
|
| 12 |
-
<h3>Usage:</h3>
|
| 13 |
-
|
| 14 |
-
<pre>
|
| 15 |
-
|
| 16 |
-
from transformers import T5Tokenizer, TFT5ForConditionalGeneration
|
| 17 |
-
|
| 18 |
-
tokenizer = T5Tokenizer.from_pretrained('SJ-Ray/Re-Punctuate')
|
| 19 |
-
model = TFT5ForConditionalGeneration.from_pretrained('SJ-Ray/Re-Punctuate')
|
| 20 |
-
|
| 21 |
-
input_text = 'the story of this brave brilliant athlete whose very being was questioned so publicly is one that still captures the imagination'
|
| 22 |
-
inputs = tokenizer.encode("punctuate: " + input_text, return_tensors="tf")
|
| 23 |
-
result = model.generate(inputs)
|
| 24 |
-
|
| 25 |
-
decoded_output = tokenizer.decode(result[0], skip_special_tokens=True)
|
| 26 |
-
print(decoded_output)
|
| 27 |
-
|
| 28 |
-
</pre>
|
| 29 |
-
<h4> Example: </h4>
|
| 30 |
-
<b>Input:</b> the story of this brave brilliant athlete whose very being was questioned so publicly is one that still captures the imagination <br>
|
| 31 |
-
<b>Output:</b> The story of this brave, brilliant athlete, whose very being was questioned so publicly, is one that still captures the imagination.
|
| 32 |
-
|
| 33 |
-
<h4> Connect on: </h4>
|
| 34 |
-
LinkedIn : www.linkedin.com/in/suraj-kumar-710382a7
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
---
|
| 4 |
+
<h2>Re-Punctuate:</h2>
|
| 5 |
+
|
| 6 |
+
Re-Punctuate is a T5 model that attempts to correct Capitalization and Punctuations in the sentences.
|
| 7 |
+
|
| 8 |
+
<h3>DataSet:</h3>
|
| 9 |
+
|
| 10 |
+
DialogSum dataset (115056 Records) was used to fine-tune the model for Punctuation and Capitalization correction.
|
| 11 |
+
|
| 12 |
+
<h3>Usage:</h3>
|
| 13 |
+
|
| 14 |
+
<pre>
|
| 15 |
+
|
| 16 |
+
from transformers import T5Tokenizer, TFT5ForConditionalGeneration
|
| 17 |
+
|
| 18 |
+
tokenizer = T5Tokenizer.from_pretrained('SJ-Ray/Re-Punctuate')
|
| 19 |
+
model = TFT5ForConditionalGeneration.from_pretrained('SJ-Ray/Re-Punctuate')
|
| 20 |
+
|
| 21 |
+
input_text = 'the story of this brave brilliant athlete whose very being was questioned so publicly is one that still captures the imagination'
|
| 22 |
+
inputs = tokenizer.encode("punctuate: " + input_text, return_tensors="tf")
|
| 23 |
+
result = model.generate(inputs)
|
| 24 |
+
|
| 25 |
+
decoded_output = tokenizer.decode(result[0], skip_special_tokens=True)
|
| 26 |
+
print(decoded_output)
|
| 27 |
+
|
| 28 |
+
</pre>
|
| 29 |
+
<h4> Example: </h4>
|
| 30 |
+
<b>Input:</b> the story of this brave brilliant athlete whose very being was questioned so publicly is one that still captures the imagination <br>
|
| 31 |
+
<b>Output:</b> The story of this brave, brilliant athlete, whose very being was questioned so publicly, is one that still captures the imagination.
|
| 32 |
+
|
| 33 |
+
<h4> Connect on: <a href="www.linkedin.com/in/suraj-kumar-710382a7">LinkedIn</a></h4>
|
|
|