t5-v1_1-base-gramatika-final-e8-b16

This model is a fine-tuned version of google/t5-v1_1-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1723
• Rouge1: 43.8331
• Rouge2: 34.7609
• Rougel: 43.5803
• Rougelsum: 43.5467
• Gen Len: 18.9287

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: 0.001
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adafactor
• lr_scheduler_type: linear
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.6434	0.37	300	0.4530	38.4418	26.1528	37.8295	37.7894	18.9434
0.5551	0.73	600	0.3368	39.883	28.2471	39.2883	39.2822	18.9345
0.4523	1.1	900	0.2959	40.3084	29.2298	39.8742	39.8747	18.9350
0.4165	1.46	1200	0.2610	41.0422	30.4902	40.6542	40.6354	18.9350
0.3196	1.83	1500	0.2292	41.6111	31.1549	41.2572	41.2477	18.9355
0.2718	2.2	1800	0.2153	41.9295	31.6902	41.5757	41.5624	18.9334
0.2446	2.56	2100	0.2055	42.2918	32.4861	42.0541	42.0135	18.9324
0.2301	2.93	2400	0.2232	42.6172	33.0243	42.3474	42.3224	18.9334
0.1997	3.29	2700	0.1859	42.8442	33.4479	42.6294	42.6121	18.9350
0.186	3.66	3000	0.1816	42.9407	33.5872	42.7248	42.7125	18.9277
0.1736	4.02	3300	0.1771	43.1994	34.0513	43.0334	42.9982	18.9308
0.1439	4.39	3600	0.1818	43.2146	33.997	43.0221	42.9893	18.9282
0.1429	4.76	3900	0.1732	43.4458	34.377	43.3072	43.26	18.9277
0.132	5.12	4200	0.1795	43.7156	34.6069	43.4982	43.481	18.9292
0.1151	5.49	4500	0.1767	43.7618	34.7345	43.5565	43.5181	18.9287
0.1127	5.85	4800	0.1723	43.8331	34.7609	43.5803	43.5467	18.9287
0.0994	6.22	5100	0.1757	43.8866	34.9216	43.641	43.6214	18.9287
0.0892	6.59	5400	0.1779	43.9415	34.9905	43.7332	43.7063	18.9292
0.0914	6.95	5700	0.1725	43.9439	35.0456	43.7419	43.7266	18.9298
0.0772	7.32	6000	0.1776	44.1132	35.3173	43.9301	43.9135	18.9287
0.0755	7.68	6300	0.1778	44.0494	35.3179	43.8797	43.8587	18.9282

Framework versions

• Transformers 4.30.1
• Pytorch 1.11.0a0+b6df043
• Datasets 2.12.0
• Tokenizers 0.13.3