Commit
·
7dfc81c
1
Parent(s):
4cacee8
Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: other
|
| 3 |
+
---
|
| 4 |
+
## Model Overview
|
| 5 |
+
|
| 6 |
+
It is a TT-compressed model of pretrained bart-base XSUM model.
|
| 7 |
+
Model was TTM-compressed with additional finetuning up to 62% (64 mln params) of original size.
|
| 8 |
+
|
| 9 |
+
- eval_loss = 2.5863
|
| 10 |
+
- eval_rouge1 = 34.5695
|
| 11 |
+
- eval_rouge2 = 12.6411
|
| 12 |
+
- eval_rougeL = 27.3062
|
| 13 |
+
- eval_rougeLsum = 27.2974
|
| 14 |
+
|
| 15 |
+
## How to use
|
| 16 |
+
|
| 17 |
+
```python
|
| 18 |
+
tokenizer = AutoTokenizer.from_pretrained("facebook/bart-base")
|
| 19 |
+
model = AutoModelForSequenceClassification.from_pretrained("s-nlp/bart-base-xsum-ttd", trust_remote_code=True)
|
| 20 |
+
```
|