Gabriel Tan
commited on
Commit
•
6243969
1
Parent(s):
386b070
Update README.md
Browse files
README.md
CHANGED
@@ -15,9 +15,22 @@ inference: false
|
|
15 |
A DialoGPT model fine-tuned on Tagalog conversational data scraped from the web. This model is an output of a research on BERT-based data augmentation for low resource languages. The base model used is DialoGPT-medium.
|
16 |
|
17 |
# Latest release: July 25, 2021
|
18 |
-
* As of the moment, the model is only able to respond based on the history of 3 previous utterances before being limited. This is a result of the
|
19 |
|
20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
# Usage
|
22 |
|
23 |
Here is an example of using Beam Search as the decoding method for our model.
|
@@ -41,5 +54,4 @@ for step in range(2):
|
|
41 |
print("DialoGPT: {}".format(tokenizer.decode(chat_history_ids[:, bot_input_ids.shape[-1]:][0], skip_special_tokens=True)))
|
42 |
```
|
43 |
|
44 |
-
|
45 |
-
To be released
|
|
|
15 |
A DialoGPT model fine-tuned on Tagalog conversational data scraped from the web. This model is an output of a research on BERT-based data augmentation for low resource languages. The base model used is DialoGPT-medium.
|
16 |
|
17 |
# Latest release: July 25, 2021
|
18 |
+
* As of the moment, the model is only able to respond based on the history of 3 previous utterances before being limited. This is a result of the scarce amount of Tagalog conversations in our dataset.
|
19 |
|
20 |
|
21 |
+
# Dataset and Scripts
|
22 |
+
The training data used was collected under the following categories:
|
23 |
+
* Food and Drinks
|
24 |
+
* Home and Garden
|
25 |
+
* Style and Fashion
|
26 |
+
* Travel and Leisure
|
27 |
+
* Visas and Immigration
|
28 |
+
* Health and Wellness
|
29 |
+
* Body and Fitness
|
30 |
+
* Small Talk
|
31 |
+
|
32 |
+
Pinoy Exchange (PEx) Conversational Dataset to be released soon.
|
33 |
+
|
34 |
# Usage
|
35 |
|
36 |
Here is an example of using Beam Search as the decoding method for our model.
|
|
|
54 |
print("DialoGPT: {}".format(tokenizer.decode(chat_history_ids[:, bot_input_ids.shape[-1]:][0], skip_special_tokens=True)))
|
55 |
```
|
56 |
|
57 |
+
|
|