bart-base-finetuned-steel-news-general

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

• Loss: 1.2510
• Rouge1: 51.0178
• Rouge2: 32.0842
• Rougel: 47.8899
• Rouge3: 0.0
• Rougelsum: 47.9633

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: 3e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 16
• 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
• lr_scheduler_warmup_steps: 500
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rouge3	Rougelsum
2.7516	0.9956	168	1.5173	46.2162	27.6554	43.4029	0.0	43.4370
1.63	1.9956	336	1.3968	47.1703	28.1491	44.5244	0.0	44.5442
1.4297	2.9956	504	1.3366	49.3461	31.7697	47.2344	0.0	47.2144
1.2575	3.9956	672	1.2632	50.3310	32.2038	47.7099	0.0	47.8190
1.0901	4.9956	840	1.2518	50.6934	32.5824	48.2610	0.0	48.3281

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1