amazon_helpfulness_classification_on_base_from_TAPT_helpfulness_pretraining

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

• Loss: 0.4874
• Accuracy: 0.8724
• F1 Macro: 0.6843

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-06
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.06
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro
0.3156	1.0	7204	0.3322	0.8666	0.5814
0.2841	2.0	14408	0.3471	0.8744	0.6461
0.274	3.0	21612	0.3581	0.8704	0.6287
0.2602	4.0	28816	0.3619	0.87	0.6849
0.2126	5.0	36020	0.5168	0.8678	0.6868
0.1674	6.0	43224	0.5960	0.8672	0.6713
0.1362	7.0	50428	0.6970	0.8684	0.6758
0.1184	8.0	57632	0.7500	0.8674	0.6715

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2