|
# Vocabulary Trimmed [lmqg/mt5-small-esquad-qa](https://huggingface.co/lmqg/mt5-small-esquad-qa): `vocabtrimmer/mt5-small-esquad-qa-trimmed-es-90000` |
|
This model is a trimmed version of [lmqg/mt5-small-esquad-qa](https://huggingface.co/lmqg/mt5-small-esquad-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-esquad-qa | vocabtrimmer/mt5-small-esquad-qa-trimmed-es-90000 | |
|
|:---------------------------|:---------------------------|:----------------------------------------------------| |
|
| parameter_size_full | 300,165,504 | 136,224,128 | |
|
| parameter_size_embedding | 256,103,424 | 92,162,048 | |
|
| vocab_size | 250,101 | 90,002 | |
|
| compression_rate_full | 100.0 | 45.38 | |
|
| compression_rate_embedding | 100.0 | 35.99 | |
|
|
|
|
|
Following table shows the parameter used to trim vocabulary. |
|
|
|
| language | dataset | dataset_column | dataset_name | dataset_split | target_vocab_size | min_frequency | |
|
|:-----------|:----------------------------|:-----------------|:---------------|:----------------|--------------------:|----------------:| |
|
| es | vocabtrimmer/mc4_validation | text | es | validation | 90000 | 2 | |