abhilash1910
commited on
Commit
•
ccb31ee
1
Parent(s):
e1609d3
README..md
Browse files
README.md
ADDED
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# DistilBERT--SQuAD-v1
|
2 |
+
|
3 |
+
Training is done on the [SQuAD](https://huggingface.co/datasets/squad) dataset. The model can be accessed via [HuggingFace](https://huggingface.co/abhilash1910/distilbert-squadv1):
|
4 |
+
|
5 |
+
|
6 |
+
## Model Specifications
|
7 |
+
|
8 |
+
We have used the following parameters:
|
9 |
+
|
10 |
+
- Training Batch Size : 512
|
11 |
+
- Learning Rate : 3e-5
|
12 |
+
- Training Epochs : 0.75
|
13 |
+
- Sequence Length : 384
|
14 |
+
- Stride : 128
|
15 |
+
|
16 |
+
|
17 |
+
## Usage Specifications
|
18 |
+
|
19 |
+
|
20 |
+
```python
|
21 |
+
|
22 |
+
from transformers import AutoModelForQuestionAnswering,AutoTokenizer,pipeline
|
23 |
+
model=AutoModelForQuestionAnswering.from_pretrained('abhilash1910/distilbert-squadv1')
|
24 |
+
tokenizer=AutoTokenizer.from_pretrained('abhilash1910/distilbert-squadv1')
|
25 |
+
nlp_QA=pipeline('question-answering',model=model,tokenizer=tokenizer)
|
26 |
+
QA_inp={
|
27 |
+
'question': 'What is the fund price of Huggingface in NYSE?',
|
28 |
+
'context': 'Huggingface Co. has a total fund price of $19.6 million dollars'
|
29 |
+
}
|
30 |
+
result=nlp_QA(QA_inp)
|
31 |
+
result
|
32 |
+
```
|
33 |
+
|
34 |
+
The result is:
|
35 |
+
|
36 |
+
```bash
|
37 |
+
|
38 |
+
{'score': 0.38547369837760925,
|
39 |
+
'start': 42,
|
40 |
+
'end': 55,
|
41 |
+
'answer': '$19.6 million'}
|
42 |
+
```
|
43 |
+
|
44 |
+
|