long_t5_test

This model is a fine-tuned version of google/long-t5-tglobal-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.3506
• Rouge Rouge1: 0.4697
• Rouge Rouge2: 0.1989
• Rouge Rougel: 0.274
• Rouge Rougelsum: 0.2736
• Gen Len: 388.0152

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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge Rouge1	Rouge Rouge2	Rouge Rougel	Rouge Rougelsum	Gen Len
No log	1.0	394	1.9389	0.0284	0.0089	0.0167	0.0165	30.2273
3.6937	2.0	788	1.4702	0.4261	0.1598	0.254	0.2539	399.0
1.8772	3.0	1182	1.4362	0.4397	0.1699	0.2592	0.2591	398.5152
1.7418	4.0	1576	1.4204	0.4434	0.1779	0.2627	0.2628	397.7374
1.7418	5.0	1970	1.4108	0.4474	0.181	0.2631	0.263	394.798
1.6623	6.0	2364	1.3932	0.4546	0.1873	0.2675	0.2673	391.8586
1.6449	7.0	2758	1.3872	0.4559	0.1882	0.2665	0.2664	393.4848
1.5757	8.0	3152	1.3814	0.458	0.1906	0.2692	0.2692	397.1061
1.5527	9.0	3546	1.3718	0.4607	0.1912	0.2705	0.2706	391.7222
1.5527	10.0	3940	1.3703	0.4649	0.194	0.2717	0.2719	393.8788
1.5302	11.0	4334	1.3621	0.4664	0.197	0.2726	0.2724	386.2071
1.5142	12.0	4728	1.3537	0.4694	0.1977	0.2731	0.2731	388.9798
1.4721	13.0	5122	1.3528	0.4652	0.1961	0.2716	0.2714	390.2828
1.4745	14.0	5516	1.3550	0.4708	0.2009	0.2742	0.2739	393.8131
1.4745	15.0	5910	1.3500	0.471	0.199	0.2742	0.2741	385.4192
1.4799	16.0	6304	1.3505	0.4725	0.2008	0.2764	0.2761	387.6364
1.4558	17.0	6698	1.3535	0.4743	0.2032	0.2765	0.2764	389.4192
1.4426	18.0	7092	1.3494	0.4743	0.2042	0.278	0.2776	386.4394
1.4426	19.0	7486	1.3513	0.4719	0.2019	0.2753	0.2752	388.6515
1.4411	20.0	7880	1.3506	0.4697	0.1989	0.274	0.2736	388.0152

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.1+cu121
• Datasets 3.0.1
• Tokenizers 0.15.1