test_trainer

This model is a fine-tuned version of mascIT/bert-tiny-ita on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5112
• Accuracy: 0.9083

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.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: 32

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	60	2.0721	0.425
No log	2.0	120	1.9418	0.4667
No log	3.0	180	1.8074	0.5833
No log	4.0	240	1.6822	0.6333
No log	5.0	300	1.5631	0.7
No log	6.0	360	1.4501	0.7333
No log	7.0	420	1.3494	0.7583
No log	8.0	480	1.2584	0.8167
1.6858	9.0	540	1.1861	0.8
1.6858	10.0	600	1.1065	0.825
1.6858	11.0	660	1.0426	0.8333
1.6858	12.0	720	0.9774	0.8333
1.6858	13.0	780	0.9263	0.825
1.6858	14.0	840	0.8681	0.8417
1.6858	15.0	900	0.8267	0.85
1.6858	16.0	960	0.7704	0.8583
0.8889	17.0	1020	0.7324	0.875
0.8889	18.0	1080	0.7094	0.8583
0.8889	19.0	1140	0.6825	0.8583
0.8889	20.0	1200	0.6484	0.8833
0.8889	21.0	1260	0.6230	0.8833
0.8889	22.0	1320	0.6056	0.875
0.8889	23.0	1380	0.5884	0.8833
0.8889	24.0	1440	0.5629	0.9
0.5022	25.0	1500	0.5537	0.8917
0.5022	26.0	1560	0.5485	0.8917
0.5022	27.0	1620	0.5411	0.8833
0.5022	28.0	1680	0.5254	0.9083
0.5022	29.0	1740	0.5198	0.9083
0.5022	30.0	1800	0.5157	0.9083
0.5022	31.0	1860	0.5107	0.9083
0.5022	32.0	1920	0.5112	0.9083

Framework versions

• Transformers 4.47.0
• Pytorch 2.3.1.post300
• Datasets 3.2.0
• Tokenizers 0.21.0