bart-base-cantonese

This model is a continue pre-train version of fnlp/bart-base-chinese on filtered Cantonese common crawl dataset with 472M tokens.

This tokenizer has extended the Bert tokenizer from fnlp/bart-base-chinese with 100 more Chinese characters commonly found in Cantonese

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1.0

Training results

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.1+cu121
• Datasets 2.14.6
• Tokenizers 0.14.1