mBART-TextSimp-LT-BatchSize8-lr5e-5

This model is a fine-tuned version of facebook/mbart-large-50 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4296
• Rouge1: 0.0605
• Rouge2: 0.0078
• Rougel: 0.0593
• Sacrebleu: 0.044
• Gen Len: 34.5776

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Sacrebleu	Gen Len
8.0008	1.0	104	7.0565	0.1958	0.1282	0.1868	7.9463	511.6945
0.3454	2.0	209	0.1874	0.6646	0.4862	0.6559	41.0808	34.5752
0.0728	3.0	313	0.0748	0.7063	0.5426	0.6984	48.033	34.5752
0.0491	4.0	418	0.0630	0.7346	0.5861	0.7248	51.6574	34.5752
0.755	5.0	522	0.7158	0.0008	0.0	0.0009	0.0	35.5752
0.4913	6.0	627	0.4653	0.0218	0.0008	0.0219	0.022	34.6134
0.4771	7.0	731	0.4525	0.0385	0.0034	0.0382	0.0308	34.926
0.4224	7.96	832	0.4296	0.0605	0.0078	0.0593	0.044	34.5776

Framework versions

• Transformers 4.33.0
• Pytorch 2.1.2+cu121
• Datasets 2.14.4
• Tokenizers 0.13.3