Abinaya Mahendiran commited on
Commit
2b02259
1 Parent(s): 36338f2

Updated app

Browse files
Files changed (2) hide show
  1. app.py +7 -9
  2. config.json +4 -1
app.py CHANGED
@@ -5,8 +5,6 @@
5
  # Install necessary libraries
6
  from transformers import AutoTokenizer, AutoModelWithLMHead, pipeline
7
  import streamlit as st
8
- from pprint import pprint
9
- import json
10
 
11
  # Read the config
12
  with open("config.json") as f:
@@ -29,22 +27,22 @@ def load_model(model_name):
29
  return model, tokenizer
30
 
31
  # Side bar
32
- img = st.sidebar.image("images/tamil_logo.jpg", width=380)
33
 
34
  # Choose the model based on selection
35
  page = st.sidebar.selectbox("Model", config["models"])
36
  data = st.sidebar.selectbox("Data", config[page])
37
 
38
  # Main page
39
- st.header("Tamil Language Demos")
40
  st.markdown(
41
  "This demo uses [GPT2 trained on Oscar dataset](https://huggingface.co/flax-community/gpt-2-tamil) "
42
  "and [GPT2 trained on Oscar & Indic Corpus dataset] (https://huggingface.co/abinayam/gpt-2-tamil) "
43
- "to show language generation"
44
  )
45
 
46
  if page == 'Text Generation' and data == 'Oscar':
47
- st.title('Tamil text generation with GPT2')
48
  st.markdown('A simple demo using gpt-2-tamil model trained on Oscar data')
49
  model, tokenizer = load_model(config[data])
50
  # Set default options
@@ -56,12 +54,12 @@ if page == 'Text Generation' and data == 'Oscar':
56
  try:
57
  with st.spinner('Generating...'):
58
  generator = pipeline('text-generation', model=model, tokenizer=tokenizer)
59
- seqs = generator(seed, max_length=max_len) # num_return_sequences=seq_num)
60
  st.write(seqs)
61
  except Exception as e:
62
  st.exception(f'Exception: {e}')
63
  elif page == 'Text Generation' and data == "Oscar + Indic Corpus":
64
- st.title('Tamil text generation with GPT2')
65
  st.markdown('A simple demo using gpt-2-tamil model trained on Oscar data')
66
  model, tokenizer = load_model(config[data])
67
  # Set default options
@@ -73,7 +71,7 @@ elif page == 'Text Generation' and data == "Oscar + Indic Corpus":
73
  try:
74
  with st.spinner('Generating...'):
75
  generator = pipeline('text-generation', model=model, tokenizer=tokenizer)
76
- seqs = generator(seed, max_length=max_len) #num_return_sequences=seq_num)
77
  st.write(seqs)
78
  except Exception as e:
79
  st.exception(f'Exception: {e}')
 
5
  # Install necessary libraries
6
  from transformers import AutoTokenizer, AutoModelWithLMHead, pipeline
7
  import streamlit as st
 
 
8
 
9
  # Read the config
10
  with open("config.json") as f:
 
27
  return model, tokenizer
28
 
29
  # Side bar
30
+ img = st.sidebar.image("images/tamil_logo.jpg", width=300)
31
 
32
  # Choose the model based on selection
33
  page = st.sidebar.selectbox("Model", config["models"])
34
  data = st.sidebar.selectbox("Data", config[page])
35
 
36
  # Main page
37
+ st.title("Tamil Language Demos")
38
  st.markdown(
39
  "This demo uses [GPT2 trained on Oscar dataset](https://huggingface.co/flax-community/gpt-2-tamil) "
40
  "and [GPT2 trained on Oscar & Indic Corpus dataset] (https://huggingface.co/abinayam/gpt-2-tamil) "
41
+ "to show language generation!"
42
  )
43
 
44
  if page == 'Text Generation' and data == 'Oscar':
45
+ st.header('Tamil text generation with GPT2')
46
  st.markdown('A simple demo using gpt-2-tamil model trained on Oscar data')
47
  model, tokenizer = load_model(config[data])
48
  # Set default options
 
54
  try:
55
  with st.spinner('Generating...'):
56
  generator = pipeline('text-generation', model=model, tokenizer=tokenizer)
57
+ seqs = generator(seed, max_length=max_len)[0]['generated_text']# num_return_sequences=seq_num)
58
  st.write(seqs)
59
  except Exception as e:
60
  st.exception(f'Exception: {e}')
61
  elif page == 'Text Generation' and data == "Oscar + Indic Corpus":
62
+ st.header('Tamil text generation with GPT2')
63
  st.markdown('A simple demo using gpt-2-tamil model trained on Oscar data')
64
  model, tokenizer = load_model(config[data])
65
  # Set default options
 
71
  try:
72
  with st.spinner('Generating...'):
73
  generator = pipeline('text-generation', model=model, tokenizer=tokenizer)
74
+ seqs = generator(seed, max_length=max_len)[0]['generated_text'] #num_return_sequences=seq_num)
75
  st.write(seqs)
76
  except Exception as e:
77
  st.exception(f'Exception: {e}')
config.json CHANGED
@@ -3,5 +3,8 @@
3
  "Text Generation": ["Oscar", "Oscar + Indic Corpus"],
4
  "Text Classification": ["News Data"],
5
  "Oscar": "flax-community/gpt-2-tamil",
6
- "Oscar + Indic Corpus": "abinayam/gpt-2-tamil"
 
 
 
7
  }
 
3
  "Text Generation": ["Oscar", "Oscar + Indic Corpus"],
4
  "Text Classification": ["News Data"],
5
  "Oscar": "flax-community/gpt-2-tamil",
6
+ "Oscar + Indic Corpus": "abinayam/gpt-2-tamil",
7
+ "examples": ["ஒரு ஊரிலே ஒரு காக்கைக்கு",
8
+ "அன்பிர்க்கும் உன்டோ அடைக்கும்",
9
+ "தெனாலி ராமன், ஒரு பெரிய விகடகவி"]
10
  }