gechim/phobert-base-v2-finetuned_60kURL

phobert-base-v2-finetuned-finetuned_60kURL

This model is a fine-tuned version of gechim/phobert-base-v2-finetuned on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3594
• Accuracy: 0.9562
• F1: 0.9563

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine_with_restarts
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.1679	1.0	704	0.1285	0.9549	0.9552
0.1111	2.0	1408	0.1405	0.9529	0.9526
0.0888	3.0	2112	0.1392	0.9592	0.9592
0.0721	4.0	2816	0.1433	0.9561	0.9564
0.059	5.0	3520	0.1563	0.9584	0.9586
0.0486	6.0	4224	0.1719	0.9549	0.9552
0.0399	7.0	4928	0.2006	0.9561	0.9563
0.0316	8.0	5632	0.2461	0.9553	0.9555
0.0269	9.0	6336	0.2424	0.9556	0.9557
0.0242	10.0	7040	0.2686	0.9543	0.9543
0.0202	11.0	7744	0.2813	0.9559	0.9559
0.0153	12.0	8448	0.2984	0.9563	0.9564
0.012	13.0	9152	0.3171	0.9553	0.9555
0.009	14.0	9856	0.3452	0.9549	0.9549
0.0088	15.0	10560	0.3415	0.9570	0.9571
0.008	16.0	11264	0.3374	0.9564	0.9564
0.0064	17.0	11968	0.3490	0.9564	0.9565
0.0054	18.0	12672	0.3598	0.9560	0.9561
0.0057	19.0	13376	0.3595	0.9559	0.9559
0.0044	20.0	14080	0.3594	0.9562	0.9563

Framework versions

• Transformers 4.41.1
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.19.1