t5-small-finetuned-xsum

This model is a fine-tuned version of t5-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.1208
• Rouge1: 36.9538
• Rouge2: 20.7839
• Rougel: 37.3871
• Rougelsum: 37.2811
• Gen Len: 18.8846

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: 2e-05
• 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: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	26	3.4305	23.5049	10.2407	21.4449	21.9379	19.0
No log	2.0	52	3.1134	24.9063	11.6883	22.8747	23.4655	18.9615
No log	3.0	78	2.8933	27.804	14.238	26.2818	26.6659	18.9615
No log	4.0	104	2.7254	32.7867	18.1634	32.9538	32.9732	18.9231
No log	5.0	130	2.6139	33.4273	18.5316	33.4935	33.5047	18.8462
No log	6.0	156	2.5310	33.3767	18.518	33.4835	33.4465	18.8462
No log	7.0	182	2.4648	34.8591	19.4604	35.0391	35.109	18.8462
No log	8.0	208	2.4072	36.9403	20.5598	37.2531	37.3462	18.8462
No log	9.0	234	2.3542	37.9145	21.2925	38.315	38.2661	18.8462
No log	10.0	260	2.3046	37.9145	21.2925	38.315	38.2661	18.8462
No log	11.0	286	2.2666	37.85	21.3084	38.1072	38.1338	18.8462
No log	12.0	312	2.2340	37.85	21.4208	38.1072	38.1338	18.8462
No log	13.0	338	2.2060	37.85	21.4208	38.1072	38.1338	18.8462
No log	14.0	364	2.1830	37.8604	21.4465	38.1781	38.1837	18.8462
No log	15.0	390	2.1646	37.8604	21.4465	38.1781	38.1837	18.8846
No log	16.0	416	2.1500	36.9538	20.7839	37.3871	37.2811	18.8846
No log	17.0	442	2.1359	36.9538	20.7839	37.3871	37.2811	18.8846
No log	18.0	468	2.1278	36.9538	20.7839	37.3871	37.2811	18.8846
No log	19.0	494	2.1227	36.9538	20.7839	37.3871	37.2811	18.8846
2.6902	20.0	520	2.1208	36.9538	20.7839	37.3871	37.2811	18.8846

Framework versions

• Transformers 4.36.0
• Pytorch 2.0.0
• Datasets 2.1.0
• Tokenizers 0.15.0