loha_fine_tuned_copa_croslo

This model is a fine-tuned version of EMBEDDIA/crosloengual-bert on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.3989
• Accuracy: 0.65
• F1: 0.6507

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: 0.003
• 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
• training_steps: 400

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.7116	1.0	50	0.6663	0.57	0.5704
0.657	2.0	100	0.6277	0.62	0.62
0.49	3.0	150	0.6608	0.67	0.6708
0.3144	4.0	200	0.7786	0.68	0.6805
0.099	5.0	250	1.1639	0.63	0.6303
0.0377	6.0	300	1.2560	0.65	0.6496
0.0216	7.0	350	1.3663	0.65	0.6503
0.0107	8.0	400	1.3989	0.65	0.6507

Framework versions

• PEFT 0.11.1
• Transformers 4.41.0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1