80K_MTL_Full_Finetuning_ST1

This model is a fine-tuned version of RonTon05/model_content_V2_test on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1942
• F1 Task1: 0.9901
• F1 Task2: 0.9349
• Acc Task1: 0.9923
• Acc Task2: 0.9777
• F1 Macro: 0.9625

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: 128
• eval_batch_size: 128
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 256
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 222
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1 Task1	F1 Task2	Acc Task1	Acc Task2	F1 Macro
0.6285	1.0	222	0.3973	0.9850	0.3644	0.9883	0.8943	0.6747
0.2553	2.0	444	0.1868	0.9882	0.7516	0.9908	0.9595	0.8699
0.1290	3.0	666	0.1485	0.9884	0.9240	0.9909	0.9724	0.9562
0.0829	4.0	888	0.1440	0.9897	0.9281	0.9920	0.9745	0.9589
0.0573	5.0	1110	0.1609	0.9892	0.9310	0.9916	0.9743	0.9601
0.0406	6.0	1332	0.1668	0.9894	0.9306	0.9918	0.9757	0.9600
0.0294	7.0	1554	0.1692	0.9902	0.9338	0.9923	0.9760	0.9620
0.0218	8.0	1776	0.1834	0.9901	0.9358	0.9923	0.9769	0.9629
0.0164	9.0	1998	0.1907	0.9904	0.9345	0.9925	0.9777	0.9624
0.0135	10.0	2220	0.1942	0.9901	0.9349	0.9923	0.9777	0.9625

Framework versions

• Transformers 5.20.2
• Pytorch 2.10.0+cu128
• Datasets 5.1.0
• Tokenizers 0.23.3