Views
No views yet
lmqg/mt5-base-dequad-ae-trimmed-50000vocabtrimmer, a tool for trimming vocabulary of language models to compress the model size.
Following table shows a summary of the trimming process.| lmqg/mt5-base-dequad-ae | lmqg/mt5-base-dequad-ae-trimmed-50000 | |
|---|---|---|
| parameter_size_full | 582,384,384 | 275,032,320 |
| parameter_size_embedding | 384,155,136 | 76,803,072 |
| vocab_size | 250,101 | 50,002 |
| compression_rate_full | 100.0 | 47.23 |
| compression_rate_embedding | 100.0 | 19.99 |
| language | dataset | dataset_column | dataset_name | dataset_split | target_vocab_size | min_frequency |
|---|---|---|---|---|---|---|
| de | vocabtrimmer/mc4_validation | text | de | validation | 50000 | 2 |