bart-caser

This model is a fine-tuned version of sn4kebyt3/ru-bart-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0552
• Bleu: 95.1367
• Meteor: 0.9722
• Chrf: 98.1048
• Gen Len: 17.4214

Model description

More information needed

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: 2e-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
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Meteor	Chrf	Gen Len
0.088	1.0	3842	0.0640	92.9548	0.9591	97.3173	17.4084
0.0319	2.0	7684	0.0516	94.7706	0.9697	97.9786	17.4139
0.0114	3.0	11526	0.0552	95.1367	0.9722	98.1048	17.4214

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1