long_t5_4

This model is a fine-tuned version of google/long-t5-tglobal-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.5697
• Rouge1: 0.5326
• Rouge2: 0.3464
• Rougel: 0.4843
• Rougelsum: 0.4839
• Gen Len: 27.734

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.4555	1.0	1000	1.7755	0.4481	0.2666	0.402	0.4019	26.852
2.2467	2.0	2000	1.7041	0.4657	0.2823	0.421	0.4205	27.304
2.1352	3.0	3000	1.6575	0.4752	0.2902	0.4295	0.4292	27.117
2.0543	4.0	4000	1.6314	0.4788	0.295	0.4348	0.4341	26.2045
2.0058	5.0	5000	1.6106	0.4856	0.3016	0.4415	0.441	26.3885
1.9365	6.0	6000	1.5924	0.4882	0.3037	0.4431	0.4425	26.048
1.9234	7.0	7000	1.5743	0.4882	0.3049	0.4435	0.443	26.207
1.8728	8.0	8000	1.5649	0.4925	0.3094	0.4479	0.4474	26.4
1.814	9.0	9000	1.5558	0.495	0.3113	0.4498	0.4495	26.383
1.8025	10.0	10000	1.5436	0.4966	0.3121	0.4517	0.4512	25.8435
1.7683	11.0	11000	1.5424	0.4987	0.3143	0.4535	0.453	25.8365
1.7299	12.0	12000	1.5308	0.4994	0.3147	0.4543	0.4537	25.718
1.7308	13.0	13000	1.5245	0.5002	0.3168	0.4554	0.4548	25.7385
1.7075	14.0	14000	1.5218	0.5028	0.3176	0.4569	0.4564	25.87
1.6969	15.0	15000	1.5171	0.5042	0.3194	0.4586	0.4583	25.7615
1.6618	16.0	16000	1.5138	0.5073	0.3216	0.4617	0.4609	25.772
1.6658	17.0	17000	1.5089	0.5051	0.3198	0.4602	0.4596	25.6465
1.6249	18.0	18000	1.5073	0.5052	0.3199	0.4604	0.4599	25.4575
1.6098	19.0	19000	1.5055	0.5068	0.321	0.4619	0.4614	26.0035
1.6018	20.0	20000	1.5015	0.5098	0.3244	0.4648	0.4644	25.4315
1.5637	21.0	21000	1.5027	0.5087	0.3243	0.4635	0.4633	26.032
1.5664	22.0	22000	1.5029	0.5118	0.3268	0.4672	0.4668	25.6305
1.561	23.0	23000	1.4968	0.5115	0.3255	0.4667	0.4661	25.7905
1.5388	24.0	24000	1.4997	0.5112	0.3259	0.4657	0.4653	26.007
1.5173	25.0	25000	1.4981	0.5129	0.3273	0.4683	0.4679	25.9415
1.5057	26.0	26000	1.4995	0.5134	0.3289	0.4692	0.4687	26.128
1.4967	27.0	27000	1.4973	0.5149	0.3308	0.4704	0.4701	25.7005
1.4755	28.0	28000	1.5033	0.5155	0.3304	0.4703	0.4699	26.4255
1.4673	29.0	29000	1.4995	0.5174	0.3319	0.4727	0.4725	25.891
1.4515	30.0	30000	1.5012	0.5158	0.3309	0.4712	0.4709	25.668
1.4502	31.0	31000	1.5021	0.518	0.3336	0.4739	0.4737	25.8405
1.4369	32.0	32000	1.4996	0.5176	0.333	0.4732	0.4729	26.093
1.4347	33.0	33000	1.5033	0.5184	0.3334	0.4731	0.4726	26.225
1.4014	34.0	34000	1.5044	0.5185	0.3333	0.4735	0.4733	26.1955
1.399	35.0	35000	1.5061	0.5192	0.3341	0.4733	0.473	26.5095
1.3941	36.0	36000	1.5067	0.5193	0.3343	0.4739	0.4735	26.2715
1.3646	37.0	37000	1.5060	0.5201	0.335	0.4753	0.4751	25.932
1.3677	38.0	38000	1.5046	0.5213	0.3354	0.4757	0.4751	26.1425
1.3623	39.0	39000	1.5084	0.5202	0.3342	0.4747	0.4743	25.9125
1.3438	40.0	40000	1.5103	0.5204	0.3356	0.4756	0.4752	26.231
1.3476	41.0	41000	1.5083	0.5203	0.3357	0.4748	0.4746	26.4745
1.3258	42.0	42000	1.5135	0.5195	0.3349	0.4744	0.474	26.265
1.3484	43.0	43000	1.5110	0.5222	0.3375	0.4762	0.476	26.4365
1.324	44.0	44000	1.5136	0.5229	0.3386	0.4781	0.4777	26.192
1.3225	45.0	45000	1.5148	0.5233	0.3377	0.477	0.4767	26.3725
1.2867	46.0	46000	1.5160	0.5224	0.3372	0.4762	0.4758	26.565
1.296	47.0	47000	1.5170	0.5224	0.3363	0.4757	0.4755	26.8325
1.2834	48.0	48000	1.5165	0.5227	0.3382	0.4772	0.477	26.5355
1.2908	49.0	49000	1.5216	0.5255	0.3391	0.4784	0.4782	26.835
1.2719	50.0	50000	1.5234	0.525	0.3392	0.4779	0.4775	26.4905
1.2768	51.0	51000	1.5257	0.5262	0.34	0.4789	0.4785	27.0845
1.2703	52.0	52000	1.5216	0.5262	0.3408	0.4798	0.4793	26.578
1.2599	53.0	53000	1.5270	0.5279	0.3409	0.4811	0.4809	26.7485
1.2502	54.0	54000	1.5250	0.5276	0.3412	0.4797	0.4794	26.8205
1.2207	55.0	55000	1.5278	0.5259	0.3408	0.4792	0.4789	26.477
1.238	56.0	56000	1.5276	0.5281	0.3423	0.4812	0.4809	26.2345
1.2199	57.0	57000	1.5303	0.5262	0.3413	0.4792	0.4788	26.818
1.2193	58.0	58000	1.5335	0.528	0.3421	0.4804	0.4802	27.0625
1.2075	59.0	59000	1.5330	0.5275	0.3405	0.4793	0.4791	27.1185
1.2096	60.0	60000	1.5401	0.5283	0.3421	0.4807	0.4805	27.2025
1.2032	61.0	61000	1.5377	0.5281	0.342	0.4806	0.4803	26.784
1.2165	62.0	62000	1.5378	0.5288	0.3423	0.4804	0.4802	27.143
1.2025	63.0	63000	1.5391	0.5275	0.3415	0.4799	0.4797	27.172
1.199	64.0	64000	1.5415	0.5303	0.3445	0.4821	0.4819	27.1665
1.1847	65.0	65000	1.5445	0.5289	0.3432	0.4815	0.4812	27.115
1.1815	66.0	66000	1.5482	0.5286	0.3428	0.4802	0.4801	27.408
1.1828	67.0	67000	1.5468	0.5299	0.3443	0.4823	0.4819	27.2485
1.1823	68.0	68000	1.5484	0.5297	0.3441	0.4813	0.4809	27.3335
1.1771	69.0	69000	1.5488	0.5305	0.3441	0.4811	0.4808	27.6115
1.1748	70.0	70000	1.5475	0.5296	0.3439	0.4814	0.4811	27.2955
1.1732	71.0	71000	1.5493	0.5304	0.3444	0.482	0.4818	27.504
1.1504	72.0	72000	1.5529	0.5305	0.3449	0.4826	0.4824	27.313
1.1497	73.0	73000	1.5528	0.5318	0.3466	0.4838	0.4835	27.463
1.1589	74.0	74000	1.5543	0.5312	0.3452	0.4826	0.4823	27.482
1.1453	75.0	75000	1.5561	0.5309	0.3447	0.4826	0.4822	27.5885
1.1451	76.0	76000	1.5577	0.5305	0.3445	0.4825	0.4822	27.3815
1.154	77.0	77000	1.5571	0.5303	0.3449	0.4828	0.4822	27.3945
1.152	78.0	78000	1.5572	0.5311	0.3456	0.4832	0.4828	27.473
1.1205	79.0	79000	1.5598	0.5317	0.3458	0.4839	0.4835	27.355
1.1376	80.0	80000	1.5619	0.5325	0.347	0.4846	0.4843	27.483
1.1391	81.0	81000	1.5614	0.5321	0.3465	0.4839	0.4835	27.7635
1.1293	82.0	82000	1.5632	0.5329	0.3472	0.4847	0.4842	27.777
1.1551	83.0	83000	1.5616	0.5323	0.3468	0.4842	0.4837	27.7005
1.1312	84.0	84000	1.5628	0.5318	0.3459	0.4835	0.4832	27.772
1.1109	85.0	85000	1.5654	0.5327	0.3469	0.4847	0.4843	27.5055
1.1371	86.0	86000	1.5653	0.534	0.3478	0.4856	0.4852	27.666
1.1355	87.0	87000	1.5642	0.5336	0.3481	0.4858	0.4855	27.617
1.1133	88.0	88000	1.5667	0.5333	0.3478	0.485	0.4847	27.725
1.1143	89.0	89000	1.5674	0.5329	0.3471	0.4849	0.4845	27.781
1.1203	90.0	90000	1.5673	0.5331	0.3474	0.4851	0.4846	27.7695
1.121	91.0	91000	1.5681	0.5333	0.3471	0.4849	0.4845	27.7595
1.0999	92.0	92000	1.5680	0.533	0.347	0.4845	0.4842	27.8525
1.1179	93.0	93000	1.5691	0.533	0.3473	0.485	0.4846	27.74
1.1057	94.0	94000	1.5688	0.5333	0.3476	0.4852	0.4847	27.6195
1.1186	95.0	95000	1.5687	0.5334	0.3474	0.4853	0.4849	27.677
1.1063	96.0	96000	1.5688	0.5329	0.3468	0.4844	0.4842	27.6925
1.0992	97.0	97000	1.5692	0.5332	0.3471	0.4849	0.4846	27.6885
1.1114	98.0	98000	1.5696	0.5328	0.3467	0.4845	0.4842	27.744
1.1085	99.0	99000	1.5697	0.5328	0.3468	0.4846	0.4842	27.744
1.101	100.0	100000	1.5697	0.5326	0.3464	0.4843	0.4839	27.734

Framework versions

• Transformers 4.45.1
• Pytorch 2.2.1
• Datasets 3.0.1
• Tokenizers 0.20.0