2023_12_22_04_32_30

This model is a fine-tuned version of google/flan-t5-large on the background_summ dataset. It achieves the following results on the evaluation set:

• Loss: 2.6624
• Rouge1: 36.2
• Rouge2: 13.9
• Rougel: 22.9
• Rougelsum: 32.7
• Bertscore Precision: 85.4
• Bertscore Recall: 86.8
• Bertscore F1: 86.1

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 1
• seed: 42
• 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	Bertscore Precision	Bertscore Recall	Bertscore F1
No log	1.0	185	1.9484	34.1	14.4	22.5	31.0	84.5	85.7	85.1
No log	2.0	370	2.0967	30.5	12.0	20.6	27.3	83.3	85.5	84.4
1.631	3.0	555	2.2225	30.8	12.0	20.7	27.2	83.5	85.8	84.6
1.631	4.0	740	2.3722	32.7	12.4	21.5	29.1	84.1	86.1	85.1
1.631	5.0	925	2.4278	34.7	13.4	22.6	31.0	84.8	86.5	85.6
1.0815	6.0	1110	2.5025	35.4	13.5	22.7	31.8	85.1	86.6	85.8
1.0815	7.0	1295	2.6083	35.9	13.9	23.0	32.4	85.2	86.8	86.0
1.0815	8.0	1480	2.6081	36.0	13.8	22.9	32.4	85.3	86.8	86.0
0.8953	9.0	1665	2.6313	36.0	13.7	22.8	32.4	85.3	86.8	86.0
0.8953	10.0	1850	2.6624	36.2	13.9	22.9	32.7	85.4	86.8	86.1

Framework versions

• Transformers 4.33.1
• Pytorch 1.13.1
• Datasets 2.14.5
• Tokenizers 0.13.3