mtc/google-mt5-large-absinth-span-3-epochs

google-mt5-large_MAX-CONTEXT-LEN-1024_MAX-GEN-LEN-256_span_absinth_faithfulness_multi_label_classification_bounded-quetzal-2024-07-15

This model is a fine-tuned version of google/mt5-large on the mtc/span_absinth_with_articles_german_faithfulness_detection_dataset dataset. It achieves the following results on the evaluation set:

• Loss: 0.1459

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
5.7665	0.1534	100	2.7347
2.3656	0.3067	200	1.6610
1.1422	0.4601	300	0.5634
0.4894	0.6135	400	0.2760
0.3222	0.7669	500	0.2368
0.3563	0.9202	600	0.1922
0.2274	1.0736	700	0.1777
0.1465	1.2270	800	0.1763
0.1499	1.3804	900	0.1732
0.1379	1.5337	1000	0.1737
0.1311	1.6871	1100	0.1615
0.1535	1.8405	1200	0.1606
0.1303	1.9939	1300	0.1637
0.0981	2.1472	1400	0.1542
0.1385	2.3006	1500	0.1311
0.124	2.4540	1600	0.1427
0.1071	2.6074	1700	0.1430
0.1127	2.7607	1800	0.1476
0.1006	2.9141	1900	0.1459

Framework versions

• Transformers 4.42.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1