ossbert-onc-unlab-from_multilingual-bs64-5epochs

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

• Loss: 0.6086

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss
1.707	0.1853	1000	1.2670
1.2452	0.3706	2000	1.0808
1.1053	0.5560	3000	0.9948
1.0305	0.7413	4000	0.9204
0.9726	0.9266	5000	0.8761
0.9261	1.1119	6000	0.8416
0.8968	1.2973	7000	0.8149
0.8626	1.4826	8000	0.7909
0.8422	1.6679	9000	0.7687
0.8218	1.8532	10000	0.7500
0.7986	2.0385	11000	0.7301
0.7784	2.2239	12000	0.7170
0.7678	2.4092	13000	0.7045
0.754	2.5945	14000	0.6946
0.7389	2.7798	15000	0.6797
0.7253	2.9652	16000	0.6723
0.718	3.1505	17000	0.6608
0.7082	3.3358	18000	0.6535
0.6995	3.5211	19000	0.6433
0.6901	3.7064	20000	0.6350
0.6847	3.8918	21000	0.6307
0.6765	4.0771	22000	0.6211
0.6676	4.2624	23000	0.6193
0.6621	4.4477	24000	0.6149
0.6595	4.6331	25000	0.6086
0.6564	4.8184	26000	0.6086

Framework versions

• Transformers 4.57.1
• Pytorch 2.9.1+cu128
• Datasets 4.0.0
• Tokenizers 0.22.1