yirmibesogluz commited on
Commit
ba2a3ae
1 Parent(s): 198feb9

Modified description

Browse files
Files changed (2) hide show
  1. app.py +36 -4
  2. images/turna-logo.png +0 -0
app.py CHANGED
@@ -4,16 +4,44 @@ from transformers import pipeline
4
  import torch
5
 
6
  DESCRIPTION="""
7
- This is the space for the Language Modeling Group at TABILAB in Computer Engineering of Bogazici University.
8
 
9
- We released the first version of our Turkish language model TURNA.
10
 
11
- This model is based on an encoder-decoder T5 architecture with 1.1B parameters.
 
 
12
 
13
- For more details, please refer to our paper.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
14
 
15
  **Note:** First inference might take time as the models are downloaded on-the-go.
16
 
 
 
17
  """
18
 
19
 
@@ -109,9 +137,13 @@ def turna(input, max_new_tokens, length_penalty,
109
 
110
  with gr.Blocks(theme="abidlabs/Lime") as demo:
111
  gr.Markdown("# TURNA 🐦")
 
 
112
  gr.Markdown(DESCRIPTION)
113
 
114
 
 
 
115
  with gr.Tab("Sentiment Analysis"):
116
  gr.Markdown("TURNA fine-tuned on sentiment analysis. Enter text to analyse sentiment and pick the model (tweets or product reviews).")
117
  with gr.Column():
 
4
  import torch
5
 
6
  DESCRIPTION="""
 
7
 
8
+ ### a Turkish encoder-decoder language model
9
 
10
+ Welcome to our Huggingface space, where you can explore the capabilities of TURNA.
11
+
12
+ **Key Features of TURNA:**
13
 
14
+ - **Powerful Architecture:** TURNA contains 1.1B parameters, and was pre-trained with an encoder-decoder architecture following the UL2 framework on 43B tokens from various domains.
15
+ - **Diverse Training Data:** Our model is trained on a varied dataset of 43 billion tokens, covering a wide array of domains.
16
+ - **Broad Applications:** TURNA is fine-tuned for a variety of generation and understanding tasks, including:
17
+ - Summarization
18
+ - Paraphrasing
19
+ - News title generation
20
+ - Sentiment classification
21
+ - Text categorization
22
+ - Named entity recognition
23
+ - Part-of-speech tagging
24
+ - Semantic textual similarity
25
+ - Natural language inference
26
+
27
+ Refer to our [paper](https://arxiv.org/abs/2401.14373) for more details.
28
+
29
+ ### Citation
30
+ ```bibtex
31
+ @misc{uludoğan2024turna,
32
+ title={TURNA: A Turkish Encoder-Decoder Language Model for Enhanced Understanding and Generation},
33
+ author={Gökçe Uludoğan and Zeynep Yirmibeşoğlu Balal and Furkan Akkurt and Melikşah Türker and Onur Güngör and Susan Üsküdarlı},
34
+ year={2024},
35
+ eprint={2401.14373},
36
+ archivePrefix={arXiv},
37
+ primaryClass={cs.CL}
38
+ }
39
+ ```
40
 
41
  **Note:** First inference might take time as the models are downloaded on-the-go.
42
 
43
+ *TURNA can generate toxic content or provide erroneous information. Double-check before usage.*
44
+
45
  """
46
 
47
 
 
137
 
138
  with gr.Blocks(theme="abidlabs/Lime") as demo:
139
  gr.Markdown("# TURNA 🐦")
140
+ gr.Image("images/turna-logo.png", scale=1)
141
+
142
  gr.Markdown(DESCRIPTION)
143
 
144
 
145
+
146
+
147
  with gr.Tab("Sentiment Analysis"):
148
  gr.Markdown("TURNA fine-tuned on sentiment analysis. Enter text to analyse sentiment and pick the model (tweets or product reviews).")
149
  with gr.Column():
images/turna-logo.png ADDED