Pannathad/xlm-roberta-base-th-product-review-sentiment-analysis

output2

This model is a fine-tuned version of xlm-roberta-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0553
• Accuracy: 0.7264
• F1: 0.7307

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: 128
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
2.5671	1.0	26	2.3849	0.2712	0.1157
2.4121	2.0	52	2.2398	0.3438	0.2029
2.2116	3.0	78	1.8608	0.4649	0.3680
1.8806	4.0	104	1.5004	0.5763	0.5189
1.5356	5.0	130	1.2657	0.6077	0.5605
1.2656	6.0	156	1.0881	0.6852	0.6578
1.0485	7.0	182	1.1436	0.6707	0.6556
0.9568	8.0	208	1.0253	0.7143	0.6974
0.813	9.0	234	0.9546	0.7070	0.6900
0.7071	10.0	260	0.9333	0.7458	0.7287
0.613	11.0	286	1.0258	0.7167	0.7038
0.5596	12.0	312	0.9554	0.7119	0.6996
0.5081	13.0	338	1.0385	0.7215	0.7147
0.4615	14.0	364	0.9769	0.7264	0.7165
0.4102	15.0	390	0.9845	0.7215	0.7213
0.3453	16.0	416	0.9315	0.7361	0.7343
0.3521	17.0	442	0.9916	0.7409	0.7439
0.2984	18.0	468	1.0486	0.7264	0.7261
0.2737	19.0	494	1.0325	0.7215	0.7239
0.2611	20.0	520	1.0210	0.7337	0.7371
0.2436	21.0	546	1.0508	0.7264	0.7283
0.2451	22.0	572	1.0487	0.7312	0.7344
0.2285	23.0	598	1.0434	0.7337	0.7366
0.2072	24.0	624	1.0530	0.7288	0.7326
0.2078	25.0	650	1.0553	0.7264	0.7307

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1