File size: 1,657 Bytes
18e1066
 
 
 
 
 
 
db8bd7d
18e1066
0bec510
18e1066
 
 
 
 
 
0bec510
18e1066
0bec510
18e1066
0bec510
18e1066
0bec510
 
 
 
 
 
 
 
 
 
 
18e1066
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0bec510
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
---
language:
- en
tags:
- generated_from_trainer
datasets:
- glue
base_model: BioLinkBERT-large
model-index:
- name: debug
  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. -->

# BioLinkBERT-large-mnli

This model is a fine-tuned version of [BioLinkBERT-large](https://huggingface.co/michiyasunaga/BioLinkBERT-large) on the GLUE [MNLI](https://huggingface.co/datasets/multi_nli) dataset.

The results are 

| **Model**              | **Dataset** | **Acc** |
|------------------------|-------------|---------|
| Roberta-large-mnli     | MNLI dev mm | 90.12   |
|                        | MNLI dev m  | 90.59   |
|                        | SNLI test   | 88.25   |
| BioLinkBERT-large      | MNLI dev mm | 33.56   |
|                        | MNLI dev m  | 33.18   |
|                        | SNLI test   | 32.66   |
| BioLinkBERT-large-mnli | MNLI dev mm | 85.19   |
|                        | MNLI dev m  | 84.96   |
|                        | SNLI test   | 78.959  |

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.5
- num_epochs: 10.0
- mixed_precision_training: Native AMP

### Framework versions

- Transformers 4.22.2
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1