nbbert_indirect_speech

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

• Accuracy: 0.8555
• Precision: 0.8632
• Recall: 0.8555
• F1: 0.8537
• Loss: 0.6652

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• 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: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Accuracy	Precision	Recall	F1	Validation Loss
No log	1.0	13	0.5937	0.6744	0.5937	0.4918	0.7937
No log	2.0	26	0.7974	0.8004	0.7974	0.7938	0.5315
No log	3.0	39	0.7321	0.7927	0.7321	0.7257	0.7020
No log	4.0	52	0.7905	0.7812	0.7905	0.7856	0.5378
No log	5.0	65	0.8130	0.8173	0.8130	0.8093	0.5648
No log	6.0	78	0.7978	0.7880	0.7978	0.7927	0.5450
No log	7.0	91	0.8313	0.8468	0.8313	0.8274	0.5858
No log	8.0	104	0.8313	0.8373	0.8313	0.8272	0.5247
No log	9.0	117	0.8294	0.8499	0.8294	0.8293	0.6528
No log	10.0	130	0.8414	0.8558	0.8414	0.8385	0.5475
No log	11.0	143	0.8472	0.8611	0.8472	0.8472	0.5893
No log	12.0	156	0.8421	0.8518	0.8421	0.8436	0.6162
No log	13.0	169	0.8382	0.8370	0.8382	0.8370	0.5678
No log	14.0	182	0.8502	0.8531	0.8502	0.8479	0.5941
No log	15.0	195	0.8521	0.8666	0.8521	0.8500	0.7085
No log	16.0	208	0.8511	0.8571	0.8511	0.8502	0.6244
No log	17.0	221	0.8548	0.8641	0.8548	0.8526	0.6625
No log	18.0	234	0.8553	0.8646	0.8553	0.8529	0.6697
No log	18.48	240	0.8555	0.8632	0.8555	0.8537	0.6652

Framework versions

• Transformers 4.48.2
• Pytorch 2.5.1+cu124
• Tokenizers 0.21.0