Super_legal_text_summarizer

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

• Loss: 1.8242
• Rouge1: 0.4168
• Rouge2: 0.1843
• Rougel: 0.26
• Rougelsum: 0.2614
• Gen Len: 126.1232

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-06
• train_batch_size: 3
• eval_batch_size: 3
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 12
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	0.9889	67	2.0691	0.3965	0.1608	0.2317	0.2325	134.8522
No log	1.9926	135	1.9581	0.4184	0.1826	0.2539	0.255	133.4433
No log	2.9963	203	1.9041	0.4129	0.1792	0.2554	0.2563	127.0591
No log	4.0	271	1.8745	0.4111	0.1769	0.2579	0.2586	126.7635
No log	4.9889	338	1.8539	0.4122	0.1754	0.258	0.2586	126.0542
No log	5.9926	406	1.8414	0.4197	0.1806	0.2603	0.2613	130.8177
No log	6.9963	474	1.8334	0.4058	0.1712	0.2532	0.2539	126.1281
1.9669	8.0	542	1.8284	0.4129	0.1818	0.2587	0.2596	125.798
1.9669	8.9889	609	1.8246	0.4129	0.1802	0.257	0.2582	126.6158
1.9669	9.8893	670	1.8242	0.4168	0.1843	0.26	0.2614	126.1232

Framework versions

• Transformers 4.40.0
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1