bart-abs-1509-0313-lr-0.0003-bs-8-maxep-10

This model is a fine-tuned version of sshleifer/distilbart-xsum-12-6 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 7.5826
• Rouge/rouge1: 0.2532
• Rouge/rouge2: 0.0528
• Rouge/rougel: 0.2067
• Rouge/rougelsum: 0.2071
• Bertscore/bertscore-precision: 0.8514
• Bertscore/bertscore-recall: 0.8621
• Bertscore/bertscore-f1: 0.8567
• Meteor: 0.2303
• Gen Len: 46.4909

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: 0.0003
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• 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	Rouge/rouge1	Rouge/rouge2	Rouge/rougel	Rouge/rougelsum	Bertscore/bertscore-precision	Bertscore/bertscore-recall	Bertscore/bertscore-f1	Meteor	Gen Len
0.6961	1.0	109	5.5084	0.2572	0.0691	0.1962	0.1962	0.8672	0.8617	0.8644	0.2158	34.0
0.6838	2.0	218	5.7494	0.2975	0.0945	0.2493	0.2495	0.8739	0.8626	0.8681	0.2433	27.0
0.5113	3.0	327	6.0212	0.2722	0.0714	0.2029	0.2031	0.8612	0.8618	0.8615	0.2582	44.0
0.4108	4.0	436	6.5957	0.2916	0.064	0.2118	0.2121	0.8678	0.8659	0.8668	0.2243	47.0
0.3585	5.0	545	6.7542	0.2554	0.0561	0.198	0.1977	0.8531	0.8633	0.8581	0.2483	42.0
0.3094	6.0	654	6.9956	0.3041	0.0711	0.2307	0.2305	0.8646	0.8658	0.8652	0.2861	42.0
0.281	7.0	763	7.1181	0.2582	0.0781	0.2156	0.2154	0.8771	0.8626	0.8697	0.1855	29.0
0.261	8.0	872	7.2717	0.3097	0.0856	0.2463	0.2464	0.8589	0.8656	0.8622	0.2246	36.0
0.2415	9.0	981	7.4446	0.2906	0.0847	0.2272	0.2274	0.8671	0.8567	0.8618	0.1991	27.0
0.2228	10.0	1090	7.5826	0.2532	0.0528	0.2067	0.2071	0.8514	0.8621	0.8567	0.2303	46.4909

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0
• Datasets 2.21.0
• Tokenizers 0.19.1