File size: 3,044 Bytes
2537f67
 
 
 
 
 
 
 
 
a1eef67
0b870e9
 
74c44e8
 
 
a1eef67
2537f67
 
 
 
 
 
 
4f4c150
0b870e9
2537f67
0b870e9
 
 
 
 
 
2537f67
0b870e9
2537f67
0b870e9
 
 
 
 
 
 
4f4c150
0b870e9
 
 
 
4f4c150
0b870e9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2537f67
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0b870e9
 
 
2537f67
 
 
 
 
 
594875a
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
---
license: mit
base_model: roberta-base
tags:
- generated_from_trainer
datasets:
- squad_v2
model-index:
- name: roberta-base-finetuned-squad2
  results: []
language:
- en
metrics:
- exact_match
- f1
pipeline_tag: question-answering
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

## Model description

RoBERTa-base fine-tuned on SQuAD 2.0 : Encoder-based Transformer Language model, pretrained with Dynamic Masking, No Next Sentence Prediction and increased Batch size compared to BERT.<br>
Suitable for Question-Answering tasks, predicts answer spans within the context provided.<br>

**Language model:** roberta-base  
**Language:** English  
**Downstream-task:** Question-Answering  
**Training data:** Train-set SQuAD 2.0  
**Evaluation data:** Evaluation-set SQuAD 2.0   
**Hardware Accelerator used**: GPU Tesla T4

## Intended uses & limitations

For Question-Answering - 

```python
!pip install transformers
from transformers import pipeline
model_checkpoint = "IProject-10/roberta-base-finetuned-squad2"
question_answerer = pipeline("question-answering", model=model_checkpoint)

context = """
🤗 Transformers is backed by the three most popular deep learning libraries — Jax, PyTorch and TensorFlow — with a seamless integration
between them. It's straightforward to train your models with one before loading them for inference with the other.
"""

question = "Which deep learning libraries back 🤗 Transformers?"
question_answerer(question=question, context=context)
```

## Results

Evaluation on SQuAD 2.0 validation dataset:

```
 exact: 79.71868946348859,
 f1: 83.049614486567,
 total: 11873,
 HasAns_exact: 78.39068825910931,
 HasAns_f1: 85.06209055313944,
 HasAns_total: 5928,
 NoAns_exact: 81.04289318755256,
 NoAns_f1: 81.04289318755256,
 NoAns_total: 5945,
 best_exact: 79.71868946348859,
 best_exact_thresh: 0.9997376203536987,
 best_f1: 83.04961448656734,
 best_f1_thresh: 0.9997376203536987,
 total_time_in_seconds: 226.245504546,
 samples_per_second: 52.47839078095801,
 latency_in_second': 0.019055462355428283
```

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3

### Training results

| Training Loss | Epoch | Step  | Validation Loss |
|:-------------:|:-----:|:-----:|:---------------:|
| 0.8921        | 1.0   | 8239  | 0.8899          |
| 0.6186        | 2.0   | 16478 | 0.8880          |
| 0.4393        | 3.0   | 24717 | 0.9785          |

This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on the squad_v2 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9785

### Framework versions

- Transformers 4.31.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.2
- Tokenizers 0.13.3