Maani commited on
Commit
c3d272e
1 Parent(s): 5b03316

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -3
README.md CHANGED
@@ -10,11 +10,13 @@ license: mit
10
 
11
  This dataset contains English-Persian translation pairs from the OPUS-100 corpus.
12
  The original corpus of OPUS-100 is movie dialogues and gets repetitive really fast,
13
- it also contains duplicates, false translations and huge chunks of unrelated data,
14
- making the dataset very noisy and almost unsuitable for any usage.
 
 
15
  We have deduplicated, denoised and cleaned the dataset, reviewed the translations,
16
  retranslated and cleared the meaning across all the entries to reach a standard version suitable for deep learning tasks.
17
- This is ~3500 first entries of the entire OPUS-100 corpus.
18
 
19
  ## Dataset Details
20
 
 
10
 
11
  This dataset contains English-Persian translation pairs from the OPUS-100 corpus.
12
  The original corpus of OPUS-100 is movie dialogues and gets repetitive really fast,
13
+ it also contains duplicates, false translations and huge chunks of unrelated data,
14
+ making the dataset very noisy and almost unsuitable for any use.
15
+
16
+ #
17
  We have deduplicated, denoised and cleaned the dataset, reviewed the translations,
18
  retranslated and cleared the meaning across all the entries to reach a standard version suitable for deep learning tasks.
19
+ This is the first 4,547 entries of the OPUS-100 corpus.
20
 
21
  ## Dataset Details
22