t5-base-asqa-cb

This model is a fine-tuned version of t5-base on the ASQA dataset. It achieves the following results on the evaluation set:

• Loss: 2.7489
• Rougelsum: 26.6134

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Rougelsum
No log	1.0	273	2.9648	23.8374
3.5538	2.0	546	2.9054	24.2701
3.5538	3.0	819	2.8744	24.4172
3.1468	4.0	1092	2.8557	24.5949
3.1468	5.0	1365	2.8400	24.7069
3.0711	6.0	1638	2.8280	24.8685
3.0711	7.0	1911	2.8191	24.9829
3.0348	8.0	2184	2.8109	25.0908
3.0348	9.0	2457	2.8038	25.2485
2.9962	10.0	2730	2.7978	25.3279
2.9635	11.0	3003	2.7920	25.4465
2.9635	12.0	3276	2.7878	25.5927
2.9328	13.0	3549	2.7833	25.6925
2.9328	14.0	3822	2.7809	25.7563
2.9126	15.0	4095	2.7773	25.8123
2.9126	16.0	4368	2.7747	25.9039
2.8878	17.0	4641	2.7719	25.9636
2.8878	18.0	4914	2.7693	26.0025
2.8744	19.0	5187	2.7673	26.0578
2.8744	20.0	5460	2.7656	26.1161
2.8579	21.0	5733	2.7629	26.1490
2.8418	22.0	6006	2.7614	26.1830
2.8418	23.0	6279	2.7604	26.2146
2.8256	24.0	6552	2.7586	26.2899
2.8256	25.0	6825	2.7586	26.2724
2.8093	26.0	7098	2.7566	26.3183
2.8093	27.0	7371	2.7551	26.3365
2.8083	28.0	7644	2.7546	26.3950
2.8083	29.0	7917	2.7537	26.4357
2.7917	30.0	8190	2.7529	26.4681
2.7917	31.0	8463	2.7526	26.5021
2.785	32.0	8736	2.7512	26.5241
2.7779	33.0	9009	2.7510	26.5361
2.7779	34.0	9282	2.7502	26.5620
2.771	35.0	9555	2.7495	26.6038
2.771	36.0	9828	2.7488	26.6161
2.7647	37.0	10101	2.7489	26.6134

Framework versions

• Transformers 4.23.0.dev0
• Pytorch 1.12.1+cu102
• Datasets 2.4.0
• Tokenizers 0.12.1