dataset-1700

This model is a fine-tuned version of facebook/mbart-large-50 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.2549
• Gen Len: 16.3667
• Rouge-1: 36.0238
• Rouge-2: 18.9307
• Rouge-l: 35.0228

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: 0.0001
• 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: polynomial
• lr_scheduler_warmup_steps: 1000
• num_epochs: 50
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Gen Len	Rouge-1	Rouge-2	Rouge-l
No log	1.0	214	3.8357	37.1467	26.95	11.9204	25.5728
No log	2.0	428	3.4294	32.64	28.5045	14.5187	26.6976
No log	3.0	642	3.3246	17.4933	27.9551	14.0601	26.9049
No log	4.0	856	3.2771	15.4	28.2521	13.6616	27.9303
No log	5.0	1070	3.1305	20.2333	34.3539	18.0221	33.5067
No log	6.0	1284	3.0782	16.9267	32.743	16.171	32.2637
No log	7.0	1498	3.0556	17.1	33.9666	17.3623	33.5188
No log	8.0	1712	3.0948	16.1067	35.7842	19.0957	35.1125
No log	9.0	1926	3.1146	16.3133	33.9124	18.8415	33.2149
No log	10.0	2140	3.1464	15.8467	35.1778	18.874	34.3936
No log	11.0	2354	3.1760	16.4467	35.6329	19.0674	34.9167
No log	12.0	2568	3.2549	16.3667	36.0238	18.9307	35.0228

Framework versions

• Transformers 4.35.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.15.2