File size: 1,967 Bytes
241bad6
f15a549
 
 
 
 
 
 
241bad6
f15a549
 
 
 
 
 
bb000fc
f15a549
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bb000fc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f15a549
 
 
 
 
 
 
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
---
tags:
- generated_from_trainer
datasets:
- squad_v2
model-index:
- name: indic-mALBERT-uncased
  results: []
---

<!-- 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. -->

# indic-mALBERT-uncased

This model is a uncased fine-tuned version of [indic-bert](https://huggingface.co/ai4bharat/indic-bert) on the squad_v2 dataset.

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 814
- num_epochs: 5.0
- mixed_precision_training: Native AMP

### Training results

| Key                     | Value                  |
|-------------------------|------------------------|
| epoch                   | 5.0                    |
| eval_HasAns_exact       | 53.54251012145749      |
| eval_HasAns_f1          | 63.501893668646716     |
| eval_HasAns_total       | 5928                   |
| eval_NoAns_exact        | 75.10513036164845      |
| eval_NoAns_f1           | 75.10513036164845      |
| eval_NoAns_total        | 5945                   |
| eval_exact              | 64.3392571380443       |
| eval_f1                 | 69.31181888888594      |
| eval_runtime            | 69.5054                |
| eval_samples            | 12000                  |
| eval_samples_per_second | 172.649                |
| eval_steps_per_second   | 5.395                  |
| eval_total              | 11873                  |

### Framework versions

- Transformers 4.29.2
- Pytorch 1.9.0+cu102
- Datasets 2.12.0
- Tokenizers 0.13.3