File size: 1,992 Bytes
55929b4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
# Vocabulary Trimmed [cardiffnlp/xlm-v-base-tweet-sentiment-en](https://huggingface.co/cardiffnlp/xlm-v-base-tweet-sentiment-en): `vocabtrimmer/xlm-v-base-tweet-sentiment-en-trimmed-en-5000` 
This model is a trimmed version of [cardiffnlp/xlm-v-base-tweet-sentiment-en](https://huggingface.co/cardiffnlp/xlm-v-base-tweet-sentiment-en) 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.

|                            | cardiffnlp/xlm-v-base-tweet-sentiment-en   | vocabtrimmer/xlm-v-base-tweet-sentiment-en-trimmed-en-5000   |
|:---------------------------|:-------------------------------------------|:-------------------------------------------------------------|
| parameter_size_full        | 778,495,491                                | 89,885,955                                                   |
| parameter_size_embedding   | 692,451,072                                | 3,841,536                                                    |
| vocab_size                 | 901,629                                    | 5,002                                                        |
| compression_rate_full      | 100.0                                      | 11.55                                                        |
| compression_rate_embedding | 100.0                                      | 0.55                                                         |


Following table shows the parameter used to trim vocabulary.

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