flan-t5-base-gen-chat_base-5

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

• Loss: 1.2877
• Rouge 1: 13.4731
• Rouge 2: 4.141
• Rouge L: 12.5428
• Avg Len: 12.3875
• Bertscore Prec: 0.8648
• Bertscore Rec: 0.8591
• Bertscore F1: 0.8616

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge 1	Rouge 2	Rouge L	Avg Len	Bertscore Prec	Bertscore Rec	Bertscore F1
3.602	0.1771	400	3.3043	6.1296	0.3906	5.6284	13.5263	0.8505	0.8478	0.8488
3.3799	0.3543	800	3.0822	6.6949	0.4835	6.216	15.4916	0.857	0.8507	0.8535
3.1711	0.5314	1200	2.8825	6.284	0.4519	5.9964	16.6215	0.8625	0.8513	0.8566
3.0298	0.7086	1600	2.7078	6.7319	0.5659	6.3339	16.9338	0.8602	0.8527	0.8562
2.8722	0.8857	2000	2.5615	7.9438	0.9859	7.4262	15.1393	0.8622	0.8538	0.8577
2.7398	1.0629	2400	2.4291	8.3717	0.9243	7.7474	14.9595	0.8601	0.854	0.8568
2.6196	1.2400	2800	2.3052	8.8458	1.117	8.2482	14.5368	0.8616	0.8544	0.8577
2.5076	1.4172	3200	2.1998	9.1369	1.2588	8.5322	14.2939	0.8626	0.855	0.8585
2.4567	1.5943	3600	2.1053	9.6663	1.4424	8.9702	12.928	0.8634	0.8547	0.8587
2.3665	1.7715	4000	2.0065	9.7718	1.4302	9.0438	13.2387	0.8622	0.8554	0.8585
2.3142	1.9486	4400	1.9322	10.2579	1.7969	9.5247	12.6887	0.8629	0.8558	0.859
2.1932	2.1258	4800	1.8453	10.4199	1.9269	9.6524	12.3864	0.8611	0.8553	0.8578
2.1476	2.3029	5200	1.7822	10.798	2.141	9.9692	12.5258	0.8617	0.8562	0.8586
2.1042	2.4801	5600	1.7094	10.9928	2.3716	10.1852	12.5231	0.8623	0.8564	0.859
2.0594	2.6572	6000	1.6480	11.5976	2.3487	10.6123	12.7434	0.8629	0.8573	0.8598
2.0017	2.8344	6400	1.5988	12.0383	2.689	11.044	12.582	0.8632	0.8577	0.8601
1.967	3.0115	6800	1.5445	11.8601	3.0155	10.9982	12.438	0.8636	0.8579	0.8604
1.9023	3.1887	7200	1.4983	12.1844	3.2931	11.3648	12.04	0.864	0.8575	0.8604
1.8604	3.3658	7600	1.4650	12.498	3.2698	11.5864	12.5668	0.8635	0.8586	0.8607
1.8345	3.5430	8000	1.4231	12.7205	3.6395	11.7643	12.4364	0.8644	0.8591	0.8614
1.8117	3.7201	8400	1.3965	12.6639	3.6483	11.824	12.4795	0.864	0.8585	0.8609
1.7999	3.8973	8800	1.3652	13.0487	3.7723	12.1147	12.4784	0.8643	0.8588	0.8612
1.7583	4.0744	9200	1.3443	13.2051	3.807	12.2615	12.4264	0.865	0.8591	0.8617
1.7608	4.2516	9600	1.3277	13.1401	3.7791	12.2312	12.4989	0.8644	0.859	0.8613
1.7111	4.4287	10000	1.3100	12.9527	3.7422	12.0599	12.4911	0.8641	0.8587	0.861
1.7137	4.6058	10400	1.2955	13.1295	4.0091	12.2326	12.5794	0.864	0.8588	0.8611
1.712	4.7830	10800	1.2899	13.511	4.1509	12.574	12.3586	0.865	0.8593	0.8618
1.6791	4.9601	11200	1.2877	13.4731	4.141	12.5428	12.3875	0.8648	0.8591	0.8616

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1