bigData_w9

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

• Loss: 1.0525
• Bleu4: 0.1309
• Rouge1: 0.4023
• Rouge2: 0.1845
• Rougel: 0.2757
• Rougelsum: 0.2753
• Gen Len: 76.4341

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: 3e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 516
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu4	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.0178	0.45	100	1.4840	0.1337	0.3875	0.1805	0.2688	0.2689	67.5509
1.142	0.89	200	1.1196	0.1296	0.3897	0.1766	0.2669	0.2667	69.3817
1.0693	1.34	300	1.0796	0.1345	0.3951	0.1818	0.2727	0.2727	70.015
1.0536	1.78	400	1.0684	0.1284	0.399	0.1839	0.2731	0.2729	77.3069
1.0084	2.23	500	1.0624	0.1287	0.3977	0.1808	0.2729	0.2729	76.7904
0.9855	2.67	600	1.0575	0.1349	0.4005	0.1843	0.2789	0.2787	72.4521
0.9812	3.12	700	1.0568	0.1303	0.4009	0.1847	0.2756	0.2752	76.1781
0.9916	3.56	800	1.0507	0.1364	0.4014	0.1853	0.279	0.2789	72.9746
0.9856	4.01	900	1.0507	0.1327	0.4003	0.1833	0.276	0.2758	74.9461
0.9747	4.45	1000	1.0519	0.1328	0.4014	0.185	0.276	0.2757	75.9461
0.9519	4.9	1100	1.0525	0.1309	0.4023	0.1845	0.2757	0.2753	76.4341

Framework versions

• Transformers 4.29.1
• Pytorch 2.0.0+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3