bart-base-instructds2

This model is a fine-tuned version of dtruong46me/bart-base-qds1 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3179
• Rouge1: 41.7719
• Rouge2: 20.1272
• Rougel: 36.3473
• Rougelsum: 38.1611
• Gen Len: 18.0

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
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
0.183	1.0	779	0.2525	42.1987	20.7169	36.597	38.6225	18.0
0.1489	2.0	1558	0.2604	41.5506	20.3178	36.2888	38.3065	18.0
0.127	3.0	2337	0.2668	41.1172	20.0439	35.6996	37.6047	18.0
0.1099	4.0	3116	0.2756	42.072	21.2877	36.7869	38.5091	18.0
0.0951	5.0	3895	0.2876	41.1625	19.5953	35.7658	37.6387	18.0
0.0838	6.0	4674	0.2964	41.59	19.9161	35.8123	37.7777	18.0
0.0738	7.0	5453	0.3064	41.3408	19.9193	35.9229	37.7199	18.0
0.066	8.0	6232	0.3109	41.7764	19.9393	36.3442	38.17	18.0
0.0603	9.0	7011	0.3141	41.747	19.9134	36.2577	38.0663	18.0
0.0558	10.0	7790	0.3179	41.7719	20.1272	36.3473	38.1611	18.0

Framework versions

• Transformers 4.36.1
• Pytorch 2.1.2
• Datasets 2.19.2
• Tokenizers 0.15.2