gagan3012 commited on
Commit
322ebac
·
1 Parent(s): 9bbcc22
Files changed (2) hide show
  1. src/data/make_dataset.py +8 -1
  2. src/models/model.py +1 -1
src/data/make_dataset.py CHANGED
@@ -1,4 +1,11 @@
1
  from datasets import load_dataset
 
2
 
3
- dataset = load_dataset('cnn_dailymail', '3.0.0')
4
 
 
 
 
 
 
 
 
 
1
  from datasets import load_dataset
2
+ import pandas as pd
3
 
 
4
 
5
+ def make_dataset(dataset='cnn_dailymail', split='train', version="3.0.0"):
6
+ """make dataset for summarisation"""
7
+ dataset = load_dataset(dataset, split=split, script_version=version)
8
+ df = pd.DataFrame()
9
+ df['input_text'] = dataset['concepts']
10
+ df['output_text'] = dataset['target']
11
+ return df
src/models/model.py CHANGED
@@ -364,7 +364,7 @@ class Summarization:
364
 
365
  def save_model(
366
  self,
367
- model_dir="models"
368
  ):
369
  """
370
  Save model to dir
 
364
 
365
  def save_model(
366
  self,
367
+ model_dir="../../models"
368
  ):
369
  """
370
  Save model to dir