asahi417's picture
commit files to HF hub
52e73f1
|
raw
history blame
1.83 kB

Vocabulary Trimmed lmqg/mbart-large-cc25-frquad-qa: vocabtrimmer/mbart-large-cc25-frquad-qa-trimmed-fr

This model is a trimmed version of lmqg/mbart-large-cc25-frquad-qa by vocabtrimmer, a tool for trimming vocabulary of language models to compress the model size. Following table shows a summary of the trimming process.

lmqg/mbart-large-cc25-frquad-qa vocabtrimmer/mbart-large-cc25-frquad-qa-trimmed-fr
parameter_size_full 610,852,864 442,588,160
parameter_size_embedding 256,028,672 87,763,968
vocab_size 250,028 85,707
compression_rate_full 100.0 72.45
compression_rate_embedding 100.0 34.28

Following table shows the parameter used to trim vocabulary.

language dataset dataset_column dataset_name dataset_split target_vocab_size min_frequency
fr vocabtrimmer/mc4_validation text fr validation 2