Edit model card
YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/model-cards#model-card-metadata)

Vocabulary Trimmed facebook/mbart-large-50: duongttr/japanese-trimmed-mbart-large

This model is a trimmed version of facebook/mbart-large-50 by vocabtrimmer, a tool for trimming vocabulary of language models to compress the model size. Following table shows a summary of the trimming process.

facebook/mbart-large-50 duongttr/japanese-trimmed-mbart-large
parameter_size_full 610,879,488 416,319,488
parameter_size_embedding 256,055,296 61,495,296
vocab_size 250,054 60,054
compression_rate_full 100.0 68.15
compression_rate_embedding 100.0 24.02

Following table shows the parameter used to trim vocabulary.

language dataset dataset_column dataset_name dataset_split target_vocab_size min_frequency
ja vocabtrimmer/mc4_validation text ja validation 60000 2
Downloads last month
3
Safetensors
Model size
416M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.