t5_base-qg-aap-test

This model is a fine-tuned version of unicamp-dl/ptt5-base-t5-vocab on the tiagoblima/preprocessed-du-qg-squadv1_pt dataset. It achieves the following results on the evaluation set:

• Loss: 0.0278

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-06
• lr_scheduler_type: linear
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	1	8.5434
No log	2.0	2	7.3013
No log	3.0	3	6.1993
No log	4.0	4	5.2898
No log	5.0	5	4.5226
No log	6.0	6	3.9202
No log	7.0	7	3.4436
No log	8.0	8	3.0408
No log	9.0	9	2.7138
No log	10.0	10	2.4436
No log	11.0	11	2.2130
No log	12.0	12	2.0190
No log	13.0	13	1.8451
No log	14.0	14	1.6746
No log	15.0	15	1.5047
No log	16.0	16	1.3376
No log	17.0	17	1.1800
No log	18.0	18	1.0434
No log	19.0	19	0.9442
No log	20.0	20	0.8739
No log	21.0	21	0.8163
No log	22.0	22	0.7629
No log	23.0	23	0.7118
No log	24.0	24	0.6618
No log	25.0	25	0.6104
No log	26.0	26	0.5597
No log	27.0	27	0.5112
No log	28.0	28	0.4657
No log	29.0	29	0.4243
No log	30.0	30	0.3873
No log	31.0	31	0.3529
No log	32.0	32	0.3209
No log	33.0	33	0.2918
No log	34.0	34	0.2667
No log	35.0	35	0.2436
No log	36.0	36	0.2215
No log	37.0	37	0.2004
No log	38.0	38	0.1808
No log	39.0	39	0.1637
No log	40.0	40	0.1484
No log	41.0	41	0.1357
No log	42.0	42	0.1252
No log	43.0	43	0.1159
No log	44.0	44	0.1079
No log	45.0	45	0.0997
No log	46.0	46	0.0922
No log	47.0	47	0.0858
No log	48.0	48	0.0802
No log	49.0	49	0.0749
No log	50.0	50	0.0703
No log	51.0	51	0.0660
No log	52.0	52	0.0626
No log	53.0	53	0.0596
No log	54.0	54	0.0573
No log	55.0	55	0.0555
No log	56.0	56	0.0539
No log	57.0	57	0.0521
No log	58.0	58	0.0506
No log	59.0	59	0.0496
No log	60.0	60	0.0485
No log	61.0	61	0.0472
No log	62.0	62	0.0460
No log	63.0	63	0.0445
No log	64.0	64	0.0432
No log	65.0	65	0.0421
No log	66.0	66	0.0409
No log	67.0	67	0.0396
No log	68.0	68	0.0385
No log	69.0	69	0.0375
No log	70.0	70	0.0365
No log	71.0	71	0.0358
No log	72.0	72	0.0350
No log	73.0	73	0.0344
No log	74.0	74	0.0338
No log	75.0	75	0.0334
No log	76.0	76	0.0329
No log	77.0	77	0.0326
No log	78.0	78	0.0321
No log	79.0	79	0.0317
No log	80.0	80	0.0314
No log	81.0	81	0.0310
No log	82.0	82	0.0306
No log	83.0	83	0.0303
No log	84.0	84	0.0299
No log	85.0	85	0.0297
No log	86.0	86	0.0294
No log	87.0	87	0.0292
No log	88.0	88	0.0290
No log	89.0	89	0.0288
No log	90.0	90	0.0287
No log	91.0	91	0.0285
No log	92.0	92	0.0284
No log	93.0	93	0.0282
No log	94.0	94	0.0281
No log	95.0	95	0.0281
No log	96.0	96	0.0280
No log	97.0	97	0.0279
No log	98.0	98	0.0279
No log	99.0	99	0.0278
0.8788	100.0	100	0.0278

Framework versions

• Transformers 4.35.2
• Pytorch 2.0.0
• Datasets 2.15.0
• Tokenizers 0.15.0