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: 0.8516
• Precision: 0.8584
• Recall: 0.8516
• F1: 0.8481
• Loss: 0.5580

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	0.5603	0.6393	0.5603	0.4202	0.8848
No log	2.0	26	0.7465	0.7387	0.7465	0.7388	0.6107
No log	3.0	39	0.7477	0.7414	0.7477	0.7442	0.7030
No log	4.0	52	0.7981	0.7980	0.7981	0.7945	0.5344
No log	5.0	65	0.8123	0.8183	0.8123	0.8087	0.4756
No log	6.0	78	0.7888	0.7790	0.7888	0.7818	0.5430
No log	7.0	91	0.8123	0.8030	0.8123	0.8075	0.5115
No log	8.0	104	0.8066	0.8012	0.8066	0.8021	0.5513
No log	9.0	117	0.8370	0.8456	0.8370	0.8371	0.4638
No log	10.0	130	0.8421	0.8377	0.8421	0.8379	0.5429
No log	11.0	143	0.8519	0.8554	0.8519	0.8496	0.4703
No log	12.0	156	0.8480	0.8428	0.8480	0.8437	0.5025
No log	13.0	169	0.8504	0.8607	0.8504	0.8499	0.5898
No log	14.0	182	0.8409	0.8342	0.8409	0.8366	0.5546
No log	15.0	195	0.8365	0.8335	0.8365	0.8339	0.5665
No log	16.0	208	0.8489	0.8503	0.8489	0.8463	0.5506
No log	17.0	221	0.8553	0.8642	0.8553	0.8521	0.5503
No log	18.0	234	0.8511	0.8577	0.8511	0.8476	0.5557
No log	18.48	240	0.8516	0.8584	0.8516	0.8481	0.5580

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Tokenizers 0.21.0