longformer-base-4096-pr

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: 0.8896
• F1 Macro: 0.6189
• Precision: 0.6241
• Recall: 0.6289
• Accuracy: 0.7711

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: 32
• seed: 42
• 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_ratio: 0.1
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1 Macro	Precision	Recall	Accuracy
No log	1.0	240	2.1101	0.3205	0.4440	0.3749	0.2669
No log	2.0	480	0.8268	0.5766	0.5856	0.6023	0.7399
1.8939	3.0	720	0.8148	0.6060	0.6099	0.6294	0.7513
1.8939	4.0	960	0.7459	0.6275	0.6266	0.6366	0.7680
0.9202	5.0	1200	0.8079	0.6122	0.6239	0.6262	0.7633
0.9202	6.0	1440	0.8896	0.6189	0.6241	0.6289	0.7711

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu128
• Datasets 4.4.1
• Tokenizers 0.22.1