Commit
•
f7b1864
1
Parent(s):
a24ea05
Update README.md
Browse files
README.md
CHANGED
@@ -6,6 +6,9 @@ The BART model was pre-trained on the CNN-DailyMail dataset, but it was re-train
|
|
6 |
|
7 |
According to huggingface, BART is a transformer encoder-encoder (seq2seq) model with a bidirectional (BERT-like) encoder and an autoregressive (GPT-like) decoder. BART is pre-trained by (1) corrupting text with an arbitrary noising function, and (2) learning a model to reconstruct the original text.
|
8 |
|
|
|
|
|
|
|
9 |
## Intended uses & limitations
|
10 |
|
11 |
This model is intended to be used for summarizing user reviews on websites.
|
|
|
6 |
|
7 |
According to huggingface, BART is a transformer encoder-encoder (seq2seq) model with a bidirectional (BERT-like) encoder and an autoregressive (GPT-like) decoder. BART is pre-trained by (1) corrupting text with an arbitrary noising function, and (2) learning a model to reconstruct the original text.
|
8 |
|
9 |
+
## Datasets
|
10 |
+
Link: [Amazon Reviews Corpus](https://huggingface.co/datasets/amazon_reviews_multi)
|
11 |
+
|
12 |
## Intended uses & limitations
|
13 |
|
14 |
This model is intended to be used for summarizing user reviews on websites.
|