Modelpublic

vocabtrimmer/xlm-roberta-base-tweet-sentiment-pt-trimmed-pt-15000

sourceHugging Faceupdated 4y agoView on Hugging Face

0likes26downloads

Model Card

Vocabulary Trimmed cardiffnlp/xlm-roberta-base-tweet-sentiment-pt: `vocabtrimmer/xlm-roberta-base-tweet-sentiment-pt-trimmed-pt-15000`

This model is a trimmed version of cardiffnlp/xlm-roberta-base-tweet-sentiment-pt by `vocabtrimmer`, a tool for trimming vocabulary of language models to compress the model size. Following table shows a summary of the trimming process.

	cardiffnlp/xlm-roberta-base-tweet-sentiment-pt	vocabtrimmer/xlm-roberta-base-tweet-sentiment-pt-trimmed-pt-15000
parametersizefull	278,045,955	97,565,955
parametersizeembedding	192,001,536	11,521,536
vocab_size	250,002	15,002
compressionratefull	100.0	35.09
compressionrateembedding	100.0	6.0

Following table shows the parameter used to trim vocabulary.

language	dataset	dataset_column	dataset_name	dataset_split	target_vocab_size	min_frequency
pt	vocabtrimmer/mc4_validation	text	pt	validation	15000	2