File size: 1,736 Bytes
1150d2e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
# Vocabulary Trimmed [lmqg/mt5-small-koquad-qa](https://huggingface.co/lmqg/mt5-small-koquad-qa): `vocabtrimmer/mt5-small-koquad-qa-trimmed-ko-5000` 
This model is a trimmed version of [lmqg/mt5-small-koquad-qa](https://huggingface.co/lmqg/mt5-small-koquad-qa) by [`vocabtrimmer`](https://github.com/asahi417/lm-vocab-trimmer), a tool for trimming vocabulary of language models to compress the model size.
Following table shows a summary of the trimming process.

|                            | lmqg/mt5-small-koquad-qa   | vocabtrimmer/mt5-small-koquad-qa-trimmed-ko-5000   |
|:---------------------------|:---------------------------|:---------------------------------------------------|
| parameter_size_full        | 300,165,504                | 49,184,128                                         |
| parameter_size_embedding   | 256,103,424                | 5,122,048                                          |
| vocab_size                 | 250,101                    | 5,002                                              |
| compression_rate_full      | 100.0                      | 16.39                                              |
| compression_rate_embedding | 100.0                      | 2.0                                                |


Following table shows the parameter used to trim vocabulary.

 | language   | dataset                     | dataset_column   | dataset_name   | dataset_split   |   target_vocab_size |   min_frequency |
|:-----------|:----------------------------|:-----------------|:---------------|:----------------|--------------------:|----------------:|
| ko         | vocabtrimmer/mc4_validation | text             | ko             | validation      |                5000 |               2 |