jackmedda/allenai-longformer-base-4096_finetuned_augmented_augmented_deepseek

This model is a fine-tuned version of allenai/longformer-base-4096 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6944
• Accuracy: 0.8529
• F1: 0.9091
• Precision: 0.8621
• Recall: 0.9615

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Use 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: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.6234	1.0	46	0.8497	0.7	0.8235	0.7	1.0
0.6942	2.0	92	0.8268	0.7	0.8235	0.7	1.0
0.4943	3.0	138	0.5841	0.7	0.8235	0.7	1.0
0.4528	4.0	184	0.5123	0.7	0.8235	0.7	1.0
0.2654	5.0	230	1.1415	0.7	0.8235	0.7	1.0
0.2954	6.0	276	0.5154	0.9	0.9333	0.875	1.0
0.4836	7.0	322	0.5581	0.9	0.9333	0.875	1.0
0.0019	8.0	368	0.7121	0.9	0.9333	0.875	1.0
0.1174	9.0	414	0.8265	0.9	0.9333	0.875	1.0
0.0005	10.0	460	0.7687	0.9	0.9333	0.875	1.0
0.0004	11.0	506	0.6626	0.9	0.9333	0.875	1.0
0.0003	12.0	552	0.5331	0.9	0.9333	0.875	1.0
0.0002	13.0	598	0.6969	0.9	0.9333	0.875	1.0
0.0003	14.0	644	0.5765	0.9	0.9333	0.875	1.0

Framework versions

• Transformers 4.48.3
• Pytorch 2.3.0+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0