flan-t5-rouge-durga-q5-clean-4c

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: 0.2052
• Rouge1: 0.4357
• Rouge2: 0.2982
• Rougel: 0.4326
• Rougelsum: 0.4327

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: 0.0001
• train_batch_size: 24
• eval_batch_size: 24
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
2.438	1.0	9	1.9807	0.2583	0.0732	0.2536	0.2534
2.6463	2.0	18	1.7018	0.2588	0.0705	0.2521	0.2525
1.7877	3.0	27	1.4796	0.2840	0.0824	0.2764	0.2765
2.0417	4.0	36	1.3040	0.3100	0.1072	0.3028	0.3027
2.0216	5.0	45	1.1630	0.3322	0.1262	0.3255	0.3246
1.7093	6.0	54	1.0289	0.3359	0.1283	0.3283	0.3285
1.6109	7.0	63	0.9288	0.3728	0.1752	0.3631	0.3628
1.3041	8.0	72	0.8358	0.3691	0.1709	0.3592	0.3593
1.3242	9.0	81	0.7609	0.3666	0.1744	0.3573	0.3579
1.0971	10.0	90	0.6803	0.3724	0.1809	0.3659	0.3663
0.7156	11.0	99	0.6153	0.3742	0.1833	0.3634	0.3637
0.8419	12.0	108	0.5537	0.3748	0.1870	0.3645	0.3655
0.8853	13.0	117	0.5012	0.3775	0.1986	0.3681	0.3687
1.0922	14.0	126	0.4396	0.3738	0.1960	0.3629	0.3634
0.8752	15.0	135	0.4022	0.3844	0.2097	0.3755	0.3762
0.8189	16.0	144	0.3810	0.4050	0.2350	0.3970	0.3975
0.639	17.0	153	0.3503	0.4039	0.2341	0.3977	0.3976
0.7971	18.0	162	0.3162	0.4082	0.2428	0.4022	0.4028
0.7211	19.0	171	0.3069	0.4174	0.2504	0.4131	0.4128
0.7633	20.0	180	0.2804	0.4204	0.2562	0.4154	0.4167
0.6475	21.0	189	0.2685	0.4308	0.2750	0.4269	0.4274
0.5642	22.0	198	0.2498	0.4232	0.2700	0.4175	0.4184
0.66	23.0	207	0.2377	0.4311	0.2832	0.4246	0.4249
0.6004	24.0	216	0.2335	0.4298	0.2868	0.4255	0.4257
0.6263	25.0	225	0.2216	0.4252	0.2806	0.4211	0.4212
0.4931	26.0	234	0.2146	0.4274	0.2858	0.4232	0.4236
0.5072	27.0	243	0.2091	0.4309	0.2862	0.4266	0.4267
0.5079	28.0	252	0.2069	0.4354	0.2969	0.4315	0.4324
0.494	29.0	261	0.2058	0.4326	0.2965	0.4290	0.4299
0.6008	30.0	270	0.2052	0.4357	0.2982	0.4326	0.4327

Framework versions

• Transformers 4.46.0
• Pytorch 2.5.0+cu121
• Datasets 3.0.2
• Tokenizers 0.20.1