miltilingual_dbert_linsearch_only_abstract

This model is a fine-tuned version of distilbert/distilbert-base-multilingual-cased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.1201
• Accuracy: 0.6505
• F1 Macro: 0.5674
• Precision Macro: 0.5715
• Recall Macro: 0.5690

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• 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: cosine
• lr_scheduler_warmup_ratio: 0.2
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	Precision Macro	Recall Macro
2.7395	1.0	1233	1.6602	0.5501	0.3447	0.3829	0.3645
1.5662	2.0	2466	1.2526	0.6228	0.5112	0.5447	0.5114
1.2526	3.0	3699	1.1599	0.6396	0.5478	0.5537	0.5551
1.1111	4.0	4932	1.1279	0.6469	0.5645	0.5619	0.5745
0.9426	5.0	6165	1.1201	0.6505	0.5674	0.5715	0.5690
0.8696	6.0	7398	1.1415	0.6462	0.5620	0.5645	0.5647
0.8271	7.0	8631	1.1486	0.6467	0.5657	0.5670	0.5667
0.7772	8.0	9864	1.1642	0.6477	0.5670	0.5644	0.5723
0.7247	9.0	11097	1.1731	0.6456	0.5644	0.5633	0.5676
0.7072	9.9922	12320	1.1731	0.6463	0.5658	0.5657	0.5677

Framework versions

• Transformers 4.50.1
• Pytorch 2.5.1+cu121
• Datasets 3.4.1
• Tokenizers 0.21.1