Update README.md
Browse files
README.md
CHANGED
@@ -13,8 +13,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
13 |
|
14 |
# Models-RoBERTa-1704501009.345538
|
15 |
|
16 |
-
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on an
|
17 |
-
It achieves the following results on the evaluation set:
|
18 |
- eval_loss: 0.2357
|
19 |
- eval_accuracy: 0.92
|
20 |
- eval_runtime: 7.4465
|
@@ -25,7 +25,15 @@ It achieves the following results on the evaluation set:
|
|
25 |
|
26 |
## Model description
|
27 |
|
28 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
|
30 |
## Intended uses & limitations
|
31 |
|
@@ -33,7 +41,12 @@ More information needed
|
|
33 |
|
34 |
## Training and evaluation data
|
35 |
|
36 |
-
|
|
|
|
|
|
|
|
|
|
|
37 |
|
38 |
## Training procedure
|
39 |
|
|
|
13 |
|
14 |
# Models-RoBERTa-1704501009.345538
|
15 |
|
16 |
+
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on an MNLI dataset.
|
17 |
+
It achieves the following results on the evaluation set (1000 instances of MNLI validation matched):
|
18 |
- eval_loss: 0.2357
|
19 |
- eval_accuracy: 0.92
|
20 |
- eval_runtime: 7.4465
|
|
|
25 |
|
26 |
## Model description
|
27 |
|
28 |
+
The baseline NLI model is a fine-tuned version of *roberta-base* for Text Classifacation
|
29 |
+
on the MNLI dataset , with entailments as label 0 and all others (neutral or contradiction)
|
30 |
+
as label 1.
|
31 |
+
|
32 |
+
Two classes:
|
33 |
+
|
34 |
+
* entailment: 0
|
35 |
+
|
36 |
+
* non-entailment: 1
|
37 |
|
38 |
## Intended uses & limitations
|
39 |
|
|
|
41 |
|
42 |
## Training and evaluation data
|
43 |
|
44 |
+
Model's performance on the validation sets:
|
45 |
+
|
46 |
+
```
|
47 |
+
MNLI: 92.07%
|
48 |
+
MNLI-mm: 92.09%
|
49 |
+
```
|
50 |
|
51 |
## Training procedure
|
52 |
|