bart-base-qds

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

• Loss: 0.1951
• Rouge1: 55.0287
• Rouge2: 38.716
• Rougel: 50.1333
• Rougelsum: 52.4893
• Gen Len: 19.8777

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 12

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
0.2775	1.0	2325	0.2104	54.7212	38.7731	49.4602	52.0184	19.8805
0.2297	2.0	4650	0.2016	55.3102	39.1181	50.2737	52.5319	19.8434
0.2158	3.0	6975	0.1987	55.5611	39.6505	50.3652	52.8286	19.8681
0.2056	4.0	9300	0.1963	54.8895	38.8942	49.7767	52.2329	19.8695
0.1982	5.0	11625	0.1955	55.3219	39.273	50.0547	52.6413	19.8695
0.192	6.0	13950	0.1953	55.4418	39.2717	50.2972	52.6911	19.8860
0.1868	7.0	16275	0.1940	55.2949	39.1619	50.3231	52.5874	19.8723
0.1829	8.0	18600	0.1943	55.4707	39.2006	50.2254	52.7875	19.875
0.1794	9.0	20925	0.1949	55.3263	38.9329	50.169	52.7384	19.8654
0.1768	10.0	23250	0.1948	54.6551	38.1892	49.6596	52.0824	19.8805
0.1751	11.0	25575	0.1952	54.9883	38.5782	49.9407	52.4031	19.8764
0.1737	12.0	27900	0.1951	55.0287	38.716	50.1333	52.4893	19.8777

Framework versions

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