bert-LoRA-reminder

This model is a fine-tuned version of dbmdz/bert-base-italian-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2139
• Accuracy: 0.9545

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6677	1.0	22	0.6283	0.7955
0.6524	2.0	44	0.6168	0.8409
0.6299	3.0	66	0.6096	0.8182
0.6258	4.0	88	0.5980	0.8636
0.6206	5.0	110	0.5849	0.8409
0.5685	6.0	132	0.5694	0.8636
0.5896	7.0	154	0.5528	0.8864
0.5636	8.0	176	0.5361	0.8636
0.5681	9.0	198	0.5217	0.8864
0.5575	10.0	220	0.4968	0.8864
0.5097	11.0	242	0.4776	0.9091
0.5001	12.0	264	0.4541	0.9091
0.4712	13.0	286	0.4269	0.9318
0.4462	14.0	308	0.4016	0.9318
0.4255	15.0	330	0.3778	0.9545
0.3943	16.0	352	0.3566	0.9545
0.3889	17.0	374	0.3358	0.9545
0.3845	18.0	396	0.3169	0.9545
0.3397	19.0	418	0.2987	0.9545
0.3677	20.0	440	0.2862	0.9545
0.3271	21.0	462	0.2729	0.9545
0.3495	22.0	484	0.2607	0.9545
0.3057	23.0	506	0.2495	0.9545
0.2621	24.0	528	0.2399	0.9545
0.2911	25.0	550	0.2314	0.9545
0.2685	26.0	572	0.2253	0.9545
0.248	27.0	594	0.2200	0.9545
0.2421	28.0	616	0.2164	0.9545
0.2688	29.0	638	0.2147	0.9545
0.2723	30.0	660	0.2139	0.9545

Framework versions

• PEFT 0.10.0
• Transformers 4.40.0
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1