T5-JSON-OM-IMP

This model is a fine-tuned version of T5-small on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.0063
Micro Precision: 0.4069
Micro Recall: 0.4681
Micro F1: 0.4353
Macro Precision: 0.4072
Macro Recall: 0.4742
Macro F1: 0.4382
Bleu: 75.6457
Rouge1: 0.7665
Rouge2: 0.5226

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: 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: 4
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Micro Precision	Micro Recall	Micro F1	Macro Precision	Macro Recall	Macro F1	Bleu	Rouge1	Rouge2
17.7048	0.1068	50	8.2715	0	0.0	0	0.0	0.0	0	0.0213	0.0194	0.0
3.5821	0.2137	100	0.1261	0	0.0	0	0.0	0.0	0	0.0022	0.0022	0.0
0.5575	0.3205	150	0.0912	0	0.0	0	0.0	0.0	0	0.0	0.0	0.0
0.1174	0.4274	200	0.0534	0	0.0	0	0.0	0.0	0	0.0	0.0	0.0
0.067	0.5342	250	0.0329	0	0.0	0	0.0	0.0	0	0.0	0.0	0.0
0.047	0.6410	300	0.0223	0	0.0	0	0.0	0.0	0	0.0000	0.0026	0.0005
0.0367	0.7479	350	0.0176	0.1667	0.0007	0.0015	0.125	0.0008	0.0015	0.0588	0.0196	0.0085
0.0305	0.8547	400	0.0151	0.3333	0.0066	0.0129	0.2500	0.0069	0.0134	0.2728	0.0411	0.0211
0.0259	0.9615	450	0.0128	0.3904	0.0536	0.0942	0.3743	0.0562	0.0978	5.1808	0.1404	0.0812
0.023	1.0684	500	0.0109	0.3549	0.1570	0.2177	0.3537	0.1627	0.2228	22.2151	0.3401	0.2093
0.0208	1.1752	550	0.0097	0.3635	0.2678	0.3084	0.3621	0.2756	0.3130	44.1481	0.5171	0.3302
0.0191	1.2821	600	0.0084	0.3442	0.3478	0.3460	0.3446	0.3517	0.3481	61.5038	0.6523	0.4297
0.017	1.3889	650	0.0080	0.3555	0.3881	0.3711	0.3555	0.3921	0.3729	67.9000	0.7144	0.4766
0.0158	1.4957	700	0.0077	0.3907	0.4299	0.4094	0.3906	0.4342	0.4112	73.1590	0.7482	0.5047
0.0151	1.6026	750	0.0073	0.3906	0.4270	0.4080	0.3905	0.4311	0.4098	74.0266	0.7572	0.5115
0.0139	1.7094	800	0.0070	0.3962	0.4453	0.4193	0.3962	0.4504	0.4216	75.0071	0.7665	0.5226
0.0135	1.8162	850	0.0069	0.4101	0.4571	0.4323	0.4106	0.4626	0.4350	75.7473	0.7674	0.5221
0.0129	1.9231	900	0.0068	0.4065	0.4563	0.4300	0.4069	0.4619	0.4327	75.3869	0.7651	0.5159
0.0125	2.0299	950	0.0067	0.3994	0.4600	0.4275	0.3995	0.4655	0.4300	75.0995	0.7644	0.5174
0.0115	2.1368	1000	0.0066	0.4059	0.4622	0.4322	0.4062	0.4678	0.4348	75.5433	0.7666	0.5209
0.0115	2.2436	1050	0.0065	0.4064	0.4637	0.4332	0.4067	0.4698	0.4360	75.5672	0.7658	0.5213
0.0114	2.3504	1100	0.0066	0.4118	0.4644	0.4366	0.4124	0.4704	0.4395	75.9796	0.7695	0.5254
0.0112	2.4573	1150	0.0065	0.4055	0.4674	0.4342	0.4057	0.4732	0.4369	75.5003	0.7654	0.5203
0.012	2.5641	1200	0.0064	0.3971	0.4585	0.4256	0.3972	0.4642	0.4281	75.0879	0.7662	0.5205
0.0107	2.6709	1250	0.0064	0.3883	0.4490	0.4165	0.3885	0.4550	0.4191	74.6715	0.7654	0.5178
0.0103	2.7778	1300	0.0064	0.3981	0.4585	0.4262	0.3983	0.4645	0.4289	75.1906	0.7673	0.5222
0.0106	2.8846	1350	0.0064	0.3985	0.4578	0.4261	0.3987	0.4639	0.4288	75.2734	0.7661	0.5196
0.0099	2.9915	1400	0.0064	0.4082	0.4681	0.4361	0.4085	0.4741	0.4389	75.6457	0.7656	0.5220
0.0101	3.0983	1450	0.0064	0.4028	0.4622	0.4305	0.4031	0.4680	0.4331	75.4024	0.7658	0.5199
0.0101	3.2051	1500	0.0064	0.4028	0.4607	0.4298	0.4031	0.4663	0.4324	75.3197	0.7648	0.5179
0.01	3.3120	1550	0.0064	0.4074	0.4666	0.4350	0.4078	0.4726	0.4378	75.5829	0.7650	0.5206
0.0095	3.4188	1600	0.0063	0.4078	0.4688	0.4362	0.4081	0.4749	0.4390	75.6771	0.7665	0.5231
0.0099	3.5256	1650	0.0063	0.4075	0.4688	0.4360	0.4078	0.4748	0.4388	75.6573	0.7662	0.5226
0.0098	3.6325	1700	0.0063	0.4069	0.4681	0.4353	0.4072	0.4742	0.4382	75.6457	0.7665	0.5226
0.0093	3.7393	1750	0.0063	0.4071	0.4681	0.4355	0.4075	0.4742	0.4383	75.6457	0.7658	0.5221
0.0095	3.8462	1800	0.0063	0.4065	0.4674	0.4348	0.4068	0.4736	0.4377	75.6341	0.7660	0.5221
0.0095	3.9530	1850	0.0063	0.4069	0.4681	0.4353	0.4072	0.4742	0.4382	75.6457	0.7665	0.5226

Framework versions

Transformers 4.47.0
Pytorch 2.5.1+cu121
Datasets 3.3.1
Tokenizers 0.21.0

lengocquangLAB/T5-JSON-OM-IMP

T5-JSON-OM-IMP

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

为此模型做种